From 9f5cf450dea680174a01bc879b053f744d635d94 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 10 Nov 2022 15:37:18 +0800 Subject: [PATCH 01/83] refactor: do some internal refactor. --- cmake/cmake.define | 2 +- include/common/tcommon.h | 6 +- include/common/tmsg.h | 32 +- source/dnode/vnode/inc/vnode.h | 2 +- source/dnode/vnode/src/meta/metaQuery.c | 4 +- source/dnode/vnode/src/tsdb/tsdbRead.c | 2 +- source/libs/executor/src/executil.c | 4 +- source/libs/function/CMakeLists.txt | 3 +- source/libs/function/inc/builtinsimpl.h | 14 +- source/libs/function/src/builtinsimpl.c | 788 +++++++++--------- .../libs/function/src/detail/tavgfunction.c | 482 +++++++++++ 11 files changed, 894 insertions(+), 445 deletions(-) create mode 100644 source/libs/function/src/detail/tavgfunction.c diff --git a/cmake/cmake.define b/cmake/cmake.define index 3f152f1f09..c7ad766a91 100644 --- a/cmake/cmake.define +++ b/cmake/cmake.define @@ -125,7 +125,7 @@ ELSE () MESSAGE("System processor ID: ${CMAKE_SYSTEM_PROCESSOR}") IF (TD_INTEL_64 OR TD_INTEL_32) - ADD_DEFINITIONS("-msse4.2") + ADD_DEFINITIONS("-msse4.2 -mavx") IF("${FMA_SUPPORT}" MATCHES "true") MESSAGE(STATUS "turn fma function support on") ADD_DEFINITIONS("-mfma") diff --git a/include/common/tcommon.h b/include/common/tcommon.h index 674bdcf171..9c1f2063a7 100644 --- a/include/common/tcommon.h +++ b/include/common/tcommon.h @@ -225,13 +225,13 @@ typedef struct SVarColAttr { // pBlockAgg->numOfNull == info.rows, all data are null // pBlockAgg->numOfNull == 0, no data are null. typedef struct SColumnInfoData { - char* pData; // the corresponding block data in memory + char* pData; // the corresponding block data in memory union { char* nullbitmap; // bitmap, one bit for each item in the list SVarColAttr varmeta; }; - SColumnInfo info; // column info - bool hasNull; // if current column data has null value. + SColumnInfo info; // column info + bool hasNull; // if current column data has null value. } SColumnInfoData; typedef struct SQueryTableDataCond { diff --git a/include/common/tmsg.h b/include/common/tmsg.h index db7d0640f6..b6f4a36ae5 100644 --- a/include/common/tmsg.h +++ b/include/common/tmsg.h @@ -643,34 +643,6 @@ int32_t tSerializeSGetUserAuthRsp(void* buf, int32_t bufLen, SGetUserAuthRsp* pR int32_t tDeserializeSGetUserAuthRsp(void* buf, int32_t bufLen, SGetUserAuthRsp* pRsp); void tFreeSGetUserAuthRsp(SGetUserAuthRsp* pRsp); -typedef struct { - int16_t lowerRelOptr; - int16_t upperRelOptr; - int16_t filterstr; // denote if current column is char(binary/nchar) - - union { - struct { - int64_t lowerBndi; - int64_t upperBndi; - }; - struct { - double lowerBndd; - double upperBndd; - }; - struct { - int64_t pz; - int64_t len; - }; - }; -} SColumnFilterInfo; - -typedef struct { - int16_t numOfFilters; - union { - int64_t placeholder; - SColumnFilterInfo* filterInfo; - }; -} SColumnFilterList; /* * for client side struct, only column id, type, bytes are necessary * But for data in vnode side, we need all the following information. @@ -681,10 +653,10 @@ typedef struct { int16_t slotId; }; - int8_t type; - int32_t bytes; uint8_t precision; uint8_t scale; + int32_t bytes; + int8_t type; } SColumnInfo; typedef struct STimeWindow { diff --git a/source/dnode/vnode/inc/vnode.h b/source/dnode/vnode/inc/vnode.h index 370103c222..0b58959822 100644 --- a/source/dnode/vnode/inc/vnode.h +++ b/source/dnode/vnode/inc/vnode.h @@ -105,7 +105,7 @@ int32_t metaGetTableTagsByUids(SMeta *pMeta, int64_t suid, SArray *uidList, int32_t metaReadNext(SMetaReader *pReader); const void *metaGetTableTagVal(void *tag, int16_t type, STagVal *tagVal); int metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName); -int metaGetTableUidByName(void *meta, char *tbName, int64_t *uid); +int metaGetTableUidByName(void *meta, char *tbName, uint64_t *uid); int metaGetTableTypeByName(void *meta, char *tbName, ETableType *tbType); bool metaIsTableExist(SMeta *pMeta, tb_uid_t uid); diff --git a/source/dnode/vnode/src/meta/metaQuery.c b/source/dnode/vnode/src/meta/metaQuery.c index 4aabd39800..32dd427d09 100644 --- a/source/dnode/vnode/src/meta/metaQuery.c +++ b/source/dnode/vnode/src/meta/metaQuery.c @@ -211,7 +211,7 @@ int metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName) { return 0; } -int metaGetTableUidByName(void *meta, char *tbName, int64_t *uid) { +int metaGetTableUidByName(void *meta, char *tbName, uint64_t *uid) { int code = 0; SMetaReader mr = {0}; metaReaderInit(&mr, (SMeta *)meta, 0); @@ -1127,7 +1127,7 @@ int32_t metaFilterTableName(SMeta *pMeta, SMetaFltParam *param, SArray *pUids) { if (valid < 0) break; char *pTableKey = (char *)pEntryKey; - int32_t cmp = (*param->filterFunc)(pTableKey, pName, pCursor->type); + cmp = (*param->filterFunc)(pTableKey, pName, pCursor->type); if (cmp == 0) { tb_uid_t tuid = *(tb_uid_t *)pEntryVal; taosArrayPush(pUids, &tuid); diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index c157faecb1..6ea270e5f4 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -535,7 +535,7 @@ static SSDataBlock* createResBlock(SQueryTableDataCond* pCond, int32_t capacity) } for (int32_t i = 0; i < pCond->numOfCols; ++i) { - SColumnInfoData colInfo = {0, {0}}; + SColumnInfoData colInfo = {0}; colInfo.info = pCond->colList[i]; blockDataAppendColInfo(pResBlock, &colInfo); } diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index d1046ff02c..f0db51dc9d 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -421,7 +421,7 @@ static SColumnInfoData* getColInfoResult(void* metaHandle, int64_t suid, SArray* } for (int32_t i = 0; i < taosArrayGetSize(ctx.cInfoList); ++i) { - SColumnInfoData colInfo = {0, {0}}; + SColumnInfoData colInfo = {0}; colInfo.info = *(SColumnInfo*)taosArrayGet(ctx.cInfoList, i); blockDataAppendColInfo(pResBlock, &colInfo); } @@ -582,7 +582,7 @@ int32_t getColInfoResultForGroupby(void* metaHandle, SNodeList* group, STableLis } for (int32_t i = 0; i < taosArrayGetSize(ctx.cInfoList); ++i) { - SColumnInfoData colInfo = {0, {0}}; + SColumnInfoData colInfo = {0}; colInfo.info = *(SColumnInfo*)taosArrayGet(ctx.cInfoList, i); blockDataAppendColInfo(pResBlock, &colInfo); } diff --git a/source/libs/function/CMakeLists.txt b/source/libs/function/CMakeLists.txt index fa241dc6ef..9d11d7b376 100644 --- a/source/libs/function/CMakeLists.txt +++ b/source/libs/function/CMakeLists.txt @@ -1,6 +1,7 @@ aux_source_directory(src FUNCTION_SRC) +aux_source_directory(src/detail FUNCTION_SRC_DETAIL) list(REMOVE_ITEM FUNCTION_SRC src/udfd.c) -add_library(function STATIC ${FUNCTION_SRC}) +add_library(function STATIC ${FUNCTION_SRC} ${FUNCTION_SRC_DETAIL}) target_include_directories( function PUBLIC diff --git a/source/libs/function/inc/builtinsimpl.h b/source/libs/function/inc/builtinsimpl.h index 89e9673b06..c0b1f62fda 100644 --- a/source/libs/function/inc/builtinsimpl.h +++ b/source/libs/function/inc/builtinsimpl.h @@ -23,6 +23,15 @@ extern "C" { #include "function.h" #include "functionMgt.h" +typedef struct SSumRes { + union { + int64_t isum; + uint64_t usum; + double dsum; + }; + int16_t type; +} SSumRes; + bool functionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo); int32_t functionFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock); int32_t functionFinalizeWithResultBuf(SqlFunctionCtx* pCtx, SSDataBlock* pBlock, char* finalResult); @@ -119,15 +128,10 @@ EFuncDataRequired lastDynDataReq(void* pRes, STimeWindow* pTimeWindow); int32_t lastRowFunction(SqlFunctionCtx* pCtx); bool getTopBotFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv); -bool getTopBotMergeFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv); bool topBotFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo); int32_t topFunction(SqlFunctionCtx* pCtx); -int32_t topFunctionMerge(SqlFunctionCtx* pCtx); int32_t bottomFunction(SqlFunctionCtx* pCtx); -int32_t bottomFunctionMerge(SqlFunctionCtx* pCtx); int32_t topBotFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock); -int32_t topBotPartialFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock); -int32_t topBotMergeFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock); int32_t topCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx); int32_t bottomCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx); int32_t getTopBotInfoSize(int64_t numOfItems); diff --git a/source/libs/function/src/builtinsimpl.c b/source/libs/function/src/builtinsimpl.c index 079e553b07..fc9c62c68f 100644 --- a/source/libs/function/src/builtinsimpl.c +++ b/source/libs/function/src/builtinsimpl.c @@ -41,21 +41,12 @@ #define HLL_BUCKET_MASK (HLL_BUCKETS - 1) #define HLL_ALPHA_INF 0.721347520444481703680 // constant for 0.5/ln(2) -typedef struct SSumRes { - union { - int64_t isum; - uint64_t usum; - double dsum; - }; - int16_t type; -} SSumRes; - -typedef struct SAvgRes { - double result; - SSumRes sum; - int64_t count; - int16_t type; // store the original input type, used in merge function -} SAvgRes; +//typedef struct SAvgRes { +// double result; +// SSumRes sum; +// int64_t count; +// int16_t type; // store the original input type, used in merge function +//} SAvgRes; typedef struct SMinmaxResInfo { bool assign; // assign the first value or not @@ -362,19 +353,19 @@ typedef struct SGroupKeyInfo { } \ } while (0) -#define LIST_AVG_N(sumT, T) \ - do { \ - T* plist = (T*)pCol->pData; \ - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { \ - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { \ - continue; \ - } \ - \ - numOfElem += 1; \ - pAvgRes->count -= 1; \ - sumT -= plist[i]; \ - } \ - } while (0) +//#define LIST_AVG_N(sumT, T) \ +// do { \ +// T* plist = (T*)pCol->pData; \ +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { \ +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { \ +// continue; \ +// } \ +// \ +// numOfElem += 1; \ +// pAvgRes->count -= 1; \ +// sumT -= plist[i]; \ +// } \ +// } while (0) #define LIST_STDDEV_SUB_N(sumT, T) \ do { \ @@ -741,374 +732,374 @@ bool getSumFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { return true; } -int32_t getAvgInfoSize() { return (int32_t)sizeof(SAvgRes); } - -bool getAvgFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { - pEnv->calcMemSize = sizeof(SAvgRes); - return true; -} - -bool avgFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo) { - if (!functionSetup(pCtx, pResultInfo)) { - return false; - } - - SAvgRes* pRes = GET_ROWCELL_INTERBUF(pResultInfo); - memset(pRes, 0, sizeof(SAvgRes)); - return true; -} - -int32_t avgFunction(SqlFunctionCtx* pCtx) { - int32_t numOfElem = 0; - - SInputColumnInfoData* pInput = &pCtx->input; - SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; - int32_t type = pInput->pData[0]->info.type; - - SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - pAvgRes->type = type; - - // computing based on the true data block - SColumnInfoData* pCol = pInput->pData[0]; - - int32_t start = pInput->startRowIndex; - int32_t numOfRows = pInput->numOfRows; - - if (IS_NULL_TYPE(type)) { - numOfElem = 0; - goto _avg_over; - } - - if (pInput->colDataAggIsSet) { - numOfElem = numOfRows - pAgg->numOfNull; - ASSERT(numOfElem >= 0); - - pAvgRes->count += numOfElem; - if (IS_SIGNED_NUMERIC_TYPE(type)) { - pAvgRes->sum.isum += pAgg->sum; - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - pAvgRes->sum.usum += pAgg->sum; - } else if (IS_FLOAT_TYPE(type)) { - pAvgRes->sum.dsum += GET_DOUBLE_VAL((const char*)&(pAgg->sum)); - } - } else { // computing based on the true data block - switch (type) { - case TSDB_DATA_TYPE_TINYINT: { - int8_t* plist = (int8_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_SMALLINT: { - int16_t* plist = (int16_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_INT: { - int32_t* plist = (int32_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_BIGINT: { - int64_t* plist = (int64_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_UTINYINT: { - uint8_t* plist = (uint8_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_USMALLINT: { - uint16_t* plist = (uint16_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_UINT: { - uint32_t* plist = (uint32_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_UBIGINT: { - uint64_t* plist = (uint64_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_FLOAT: { - float* plist = (float*)pCol->pData; -// float val = 0; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.dsum += plist[i]; - } -// pAvgRes->sum.dsum = val; - break; - } - - case TSDB_DATA_TYPE_DOUBLE: { - double* plist = (double*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.dsum += plist[i]; - } - break; - } - - default: - break; - } - } - -_avg_over: - // data in the check operation are all null, not output - SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); - return TSDB_CODE_SUCCESS; -} - -static void avgTransferInfo(SAvgRes* pInput, SAvgRes* pOutput) { - pOutput->type = pInput->type; - if (IS_SIGNED_NUMERIC_TYPE(pOutput->type)) { - pOutput->sum.isum += pInput->sum.isum; - } else if (IS_UNSIGNED_NUMERIC_TYPE(pOutput->type)) { - pOutput->sum.usum += pInput->sum.usum; - } else { - pOutput->sum.dsum += pInput->sum.dsum; - } - - pOutput->count += pInput->count; - - return; -} - -int32_t avgFunctionMerge(SqlFunctionCtx* pCtx) { - SInputColumnInfoData* pInput = &pCtx->input; - SColumnInfoData* pCol = pInput->pData[0]; - ASSERT(pCol->info.type == TSDB_DATA_TYPE_BINARY); - - SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - - int32_t start = pInput->startRowIndex; - - for (int32_t i = start; i < start + pInput->numOfRows; ++i) { - char* data = colDataGetData(pCol, i); - SAvgRes* pInputInfo = (SAvgRes*)varDataVal(data); - avgTransferInfo(pInputInfo, pInfo); - } - - SET_VAL(GET_RES_INFO(pCtx), 1, 1); - - return TSDB_CODE_SUCCESS; -} - -int32_t avgInvertFunction(SqlFunctionCtx* pCtx) { - int32_t numOfElem = 0; - - // Only the pre-computing information loaded and actual data does not loaded - SInputColumnInfoData* pInput = &pCtx->input; - int32_t type = pInput->pData[0]->info.type; - - SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - - // computing based on the true data block - SColumnInfoData* pCol = pInput->pData[0]; - - int32_t start = pInput->startRowIndex; - int32_t numOfRows = pInput->numOfRows; - - switch (type) { - case TSDB_DATA_TYPE_TINYINT: { - LIST_AVG_N(pAvgRes->sum.isum, int8_t); - break; - } - case TSDB_DATA_TYPE_SMALLINT: { - LIST_AVG_N(pAvgRes->sum.isum, int16_t); - break; - } - case TSDB_DATA_TYPE_INT: { - LIST_AVG_N(pAvgRes->sum.isum, int32_t); - break; - } - case TSDB_DATA_TYPE_BIGINT: { - LIST_AVG_N(pAvgRes->sum.isum, int64_t); - break; - } - case TSDB_DATA_TYPE_UTINYINT: { - LIST_AVG_N(pAvgRes->sum.usum, uint8_t); - break; - } - case TSDB_DATA_TYPE_USMALLINT: { - LIST_AVG_N(pAvgRes->sum.usum, uint16_t); - break; - } - case TSDB_DATA_TYPE_UINT: { - LIST_AVG_N(pAvgRes->sum.usum, uint32_t); - break; - } - case TSDB_DATA_TYPE_UBIGINT: { - LIST_AVG_N(pAvgRes->sum.usum, uint64_t); - break; - } - case TSDB_DATA_TYPE_FLOAT: { - LIST_AVG_N(pAvgRes->sum.dsum, float); - break; - } - case TSDB_DATA_TYPE_DOUBLE: { - LIST_AVG_N(pAvgRes->sum.dsum, double); - break; - } - default: - break; - } - - // data in the check operation are all null, not output - SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); - return TSDB_CODE_SUCCESS; -} - -int32_t avgCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx) { - SResultRowEntryInfo* pDResInfo = GET_RES_INFO(pDestCtx); - SAvgRes* pDBuf = GET_ROWCELL_INTERBUF(pDResInfo); - - SResultRowEntryInfo* pSResInfo = GET_RES_INFO(pSourceCtx); - SAvgRes* pSBuf = GET_ROWCELL_INTERBUF(pSResInfo); - int16_t type = pDBuf->type == TSDB_DATA_TYPE_NULL ? pSBuf->type : pDBuf->type; - - if (IS_SIGNED_NUMERIC_TYPE(type)) { - pDBuf->sum.isum += pSBuf->sum.isum; - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - pDBuf->sum.usum += pSBuf->sum.usum; - } else { - pDBuf->sum.dsum += pSBuf->sum.dsum; - } - pDBuf->count += pSBuf->count; - - return TSDB_CODE_SUCCESS; -} - -int32_t avgFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { - SInputColumnInfoData* pInput = &pCtx->input; - - SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - int32_t type = pAvgRes->type; - - if (IS_SIGNED_NUMERIC_TYPE(type)) { - pAvgRes->result = pAvgRes->sum.isum / ((double)pAvgRes->count); - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - pAvgRes->result = pAvgRes->sum.usum / ((double)pAvgRes->count); - } else { - pAvgRes->result = pAvgRes->sum.dsum / ((double)pAvgRes->count); - } - - // check for overflow - if (isinf(pAvgRes->result) || isnan(pAvgRes->result)) { - GET_RES_INFO(pCtx)->numOfRes = 0; - } - - return functionFinalize(pCtx, pBlock); -} - -int32_t avgPartialFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { - SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); - SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - int32_t resultBytes = getAvgInfoSize(); - char* res = taosMemoryCalloc(resultBytes + VARSTR_HEADER_SIZE, sizeof(char)); - - memcpy(varDataVal(res), pInfo, resultBytes); - varDataSetLen(res, resultBytes); - - int32_t slotId = pCtx->pExpr->base.resSchema.slotId; - SColumnInfoData* pCol = taosArrayGet(pBlock->pDataBlock, slotId); - - colDataAppend(pCol, pBlock->info.rows, res, false); - - taosMemoryFree(res); - return pResInfo->numOfRes; -} +//int32_t getAvgInfoSize() { return (int32_t)sizeof(SAvgRes); } +// +//bool getAvgFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { +// pEnv->calcMemSize = sizeof(SAvgRes); +// return true; +//} +// +//bool avgFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo) { +// if (!functionSetup(pCtx, pResultInfo)) { +// return false; +// } +// +// SAvgRes* pRes = GET_ROWCELL_INTERBUF(pResultInfo); +// memset(pRes, 0, sizeof(SAvgRes)); +// return true; +//} + +//int32_t avgFunction(SqlFunctionCtx* pCtx) { +// int32_t numOfElem = 0; +// +// SInputColumnInfoData* pInput = &pCtx->input; +// SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; +// int32_t type = pInput->pData[0]->info.type; +// +// SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); +// pAvgRes->type = type; +// +// // computing based on the true data block +// SColumnInfoData* pCol = pInput->pData[0]; +// +// int32_t start = pInput->startRowIndex; +// int32_t numOfRows = pInput->numOfRows; +// +// if (IS_NULL_TYPE(type)) { +// numOfElem = 0; +// goto _avg_over; +// } +// +// if (pInput->colDataAggIsSet) { +// numOfElem = numOfRows - pAgg->numOfNull; +// ASSERT(numOfElem >= 0); +// +// pAvgRes->count += numOfElem; +// if (IS_SIGNED_NUMERIC_TYPE(type)) { +// pAvgRes->sum.isum += pAgg->sum; +// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { +// pAvgRes->sum.usum += pAgg->sum; +// } else if (IS_FLOAT_TYPE(type)) { +// pAvgRes->sum.dsum += GET_DOUBLE_VAL((const char*)&(pAgg->sum)); +// } +// } else { // computing based on the true data block +// switch (type) { +// case TSDB_DATA_TYPE_TINYINT: { +// int8_t* plist = (int8_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.isum += plist[i]; +// } +// +// break; +// } +// +// case TSDB_DATA_TYPE_SMALLINT: { +// int16_t* plist = (int16_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.isum += plist[i]; +// } +// break; +// } +// +// case TSDB_DATA_TYPE_INT: { +// int32_t* plist = (int32_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.isum += plist[i]; +// } +// +// break; +// } +// +// case TSDB_DATA_TYPE_BIGINT: { +// int64_t* plist = (int64_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.isum += plist[i]; +// } +// break; +// } +// +// case TSDB_DATA_TYPE_UTINYINT: { +// uint8_t* plist = (uint8_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.usum += plist[i]; +// } +// +// break; +// } +// +// case TSDB_DATA_TYPE_USMALLINT: { +// uint16_t* plist = (uint16_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.usum += plist[i]; +// } +// break; +// } +// +// case TSDB_DATA_TYPE_UINT: { +// uint32_t* plist = (uint32_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.usum += plist[i]; +// } +// +// break; +// } +// +// case TSDB_DATA_TYPE_UBIGINT: { +// uint64_t* plist = (uint64_t*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.usum += plist[i]; +// } +// break; +// } +// +// case TSDB_DATA_TYPE_FLOAT: { +// float* plist = (float*)pCol->pData; +//// float val = 0; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.dsum += plist[i]; +// } +//// pAvgRes->sum.dsum = val; +// break; +// } +// +// case TSDB_DATA_TYPE_DOUBLE: { +// double* plist = (double*)pCol->pData; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.dsum += plist[i]; +// } +// break; +// } +// +// default: +// break; +// } +// } +// +//_avg_over: +// // data in the check operation are all null, not output +// SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); +// return TSDB_CODE_SUCCESS; +//} + +//static void avgTransferInfo(SAvgRes* pInput, SAvgRes* pOutput) { +// pOutput->type = pInput->type; +// if (IS_SIGNED_NUMERIC_TYPE(pOutput->type)) { +// pOutput->sum.isum += pInput->sum.isum; +// } else if (IS_UNSIGNED_NUMERIC_TYPE(pOutput->type)) { +// pOutput->sum.usum += pInput->sum.usum; +// } else { +// pOutput->sum.dsum += pInput->sum.dsum; +// } +// +// pOutput->count += pInput->count; +// +// return; +//} +// +//int32_t avgFunctionMerge(SqlFunctionCtx* pCtx) { +// SInputColumnInfoData* pInput = &pCtx->input; +// SColumnInfoData* pCol = pInput->pData[0]; +// ASSERT(pCol->info.type == TSDB_DATA_TYPE_BINARY); +// +// SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); +// +// int32_t start = pInput->startRowIndex; +// +// for (int32_t i = start; i < start + pInput->numOfRows; ++i) { +// char* data = colDataGetData(pCol, i); +// SAvgRes* pInputInfo = (SAvgRes*)varDataVal(data); +// avgTransferInfo(pInputInfo, pInfo); +// } +// +// SET_VAL(GET_RES_INFO(pCtx), 1, 1); +// +// return TSDB_CODE_SUCCESS; +//} +// +//int32_t avgInvertFunction(SqlFunctionCtx* pCtx) { +// int32_t numOfElem = 0; +// +// // Only the pre-computing information loaded and actual data does not loaded +// SInputColumnInfoData* pInput = &pCtx->input; +// int32_t type = pInput->pData[0]->info.type; +// +// SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); +// +// // computing based on the true data block +// SColumnInfoData* pCol = pInput->pData[0]; +// +// int32_t start = pInput->startRowIndex; +// int32_t numOfRows = pInput->numOfRows; +// +// switch (type) { +// case TSDB_DATA_TYPE_TINYINT: { +// LIST_AVG_N(pAvgRes->sum.isum, int8_t); +// break; +// } +// case TSDB_DATA_TYPE_SMALLINT: { +// LIST_AVG_N(pAvgRes->sum.isum, int16_t); +// break; +// } +// case TSDB_DATA_TYPE_INT: { +// LIST_AVG_N(pAvgRes->sum.isum, int32_t); +// break; +// } +// case TSDB_DATA_TYPE_BIGINT: { +// LIST_AVG_N(pAvgRes->sum.isum, int64_t); +// break; +// } +// case TSDB_DATA_TYPE_UTINYINT: { +// LIST_AVG_N(pAvgRes->sum.usum, uint8_t); +// break; +// } +// case TSDB_DATA_TYPE_USMALLINT: { +// LIST_AVG_N(pAvgRes->sum.usum, uint16_t); +// break; +// } +// case TSDB_DATA_TYPE_UINT: { +// LIST_AVG_N(pAvgRes->sum.usum, uint32_t); +// break; +// } +// case TSDB_DATA_TYPE_UBIGINT: { +// LIST_AVG_N(pAvgRes->sum.usum, uint64_t); +// break; +// } +// case TSDB_DATA_TYPE_FLOAT: { +// LIST_AVG_N(pAvgRes->sum.dsum, float); +// break; +// } +// case TSDB_DATA_TYPE_DOUBLE: { +// LIST_AVG_N(pAvgRes->sum.dsum, double); +// break; +// } +// default: +// break; +// } +// +// // data in the check operation are all null, not output +// SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); +// return TSDB_CODE_SUCCESS; +//} +// +//int32_t avgCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx) { +// SResultRowEntryInfo* pDResInfo = GET_RES_INFO(pDestCtx); +// SAvgRes* pDBuf = GET_ROWCELL_INTERBUF(pDResInfo); +// +// SResultRowEntryInfo* pSResInfo = GET_RES_INFO(pSourceCtx); +// SAvgRes* pSBuf = GET_ROWCELL_INTERBUF(pSResInfo); +// int16_t type = pDBuf->type == TSDB_DATA_TYPE_NULL ? pSBuf->type : pDBuf->type; +// +// if (IS_SIGNED_NUMERIC_TYPE(type)) { +// pDBuf->sum.isum += pSBuf->sum.isum; +// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { +// pDBuf->sum.usum += pSBuf->sum.usum; +// } else { +// pDBuf->sum.dsum += pSBuf->sum.dsum; +// } +// pDBuf->count += pSBuf->count; +// +// return TSDB_CODE_SUCCESS; +//} +// +//int32_t avgFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { +// SInputColumnInfoData* pInput = &pCtx->input; +// +// SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); +// int32_t type = pAvgRes->type; +// +// if (IS_SIGNED_NUMERIC_TYPE(type)) { +// pAvgRes->result = pAvgRes->sum.isum / ((double)pAvgRes->count); +// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { +// pAvgRes->result = pAvgRes->sum.usum / ((double)pAvgRes->count); +// } else { +// pAvgRes->result = pAvgRes->sum.dsum / ((double)pAvgRes->count); +// } +// +// // check for overflow +// if (isinf(pAvgRes->result) || isnan(pAvgRes->result)) { +// GET_RES_INFO(pCtx)->numOfRes = 0; +// } +// +// return functionFinalize(pCtx, pBlock); +//} +// +//int32_t avgPartialFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { +// SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); +// SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); +// int32_t resultBytes = getAvgInfoSize(); +// char* res = taosMemoryCalloc(resultBytes + VARSTR_HEADER_SIZE, sizeof(char)); +// +// memcpy(varDataVal(res), pInfo, resultBytes); +// varDataSetLen(res, resultBytes); +// +// int32_t slotId = pCtx->pExpr->base.resSchema.slotId; +// SColumnInfoData* pCol = taosArrayGet(pBlock->pDataBlock, slotId); +// +// colDataAppend(pCol, pBlock->info.rows, res, false); +// +// taosMemoryFree(res); +// return pResInfo->numOfRes; +//} EFuncDataRequired statisDataRequired(SFunctionNode* pFunc, STimeWindow* pTimeWindow) { return FUNC_DATA_REQUIRED_SMA_LOAD; @@ -3117,8 +3108,7 @@ int32_t lastFunction(SqlFunctionCtx* pCtx) { int32_t round = pInput->numOfRows >> 2; int32_t reminder = pInput->numOfRows & 0x03; - int32_t tick = 0; - for (int32_t i = pInput->startRowIndex; tick < round; i += 4, tick += 1) { + for (int32_t i = pInput->startRowIndex, tick = 0; tick < round; i += 4, tick += 1) { int64_t cts = pts[i]; int32_t chosen = i; @@ -3153,7 +3143,7 @@ int32_t lastFunction(SqlFunctionCtx* pCtx) { } } else { for (int32_t i = pInput->startRowIndex; i < pInput->startRowIndex + pInput->numOfRows; ++i) { - if (pInputCol->hasNull && colDataIsNull(pInputCol, pInput->totalRows, i, pColAgg)) { + if (colDataIsNull(pInputCol, pInput->totalRows, i, pColAgg)) { continue; } diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c new file mode 100644 index 0000000000..431e169346 --- /dev/null +++ b/source/libs/function/src/detail/tavgfunction.c @@ -0,0 +1,482 @@ +/* + * Copyright (c) 2019 TAOS Data, Inc. + * + * This program is free software: you can use, redistribute, and/or modify + * it under the terms of the GNU Affero General Public License, version 3 + * or later ("AGPL"), as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. + * + * You should have received a copy of the GNU Affero General Public License + * along with this program. If not, see . + */ + +#include +#include "builtinsimpl.h" +#include "function.h" +#include "tdatablock.h" +#include "tfunctionInt.h" +#include "tglobal.h" + +#define SET_VAL(_info, numOfElem, res) \ + do { \ + if ((numOfElem) <= 0) { \ + break; \ + } \ + (_info)->numOfRes = (res); \ + } while (0) + +#define LIST_AVG_N(sumT, T) \ + do { \ + T* plist = (T*)pCol->pData; \ + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { \ + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { \ + continue; \ + } \ + \ + numOfElem += 1; \ + pAvgRes->count -= 1; \ + sumT -= plist[i]; \ + } \ + } while (0) + +typedef struct SAvgRes { + double result; + SSumRes sum; + int64_t count; + int16_t type; // store the original input type, used in merge function +} SAvgRes; + +static int32_t handleFloatCols(const SColumnInfoData* pCol, const SInputColumnInfoData* pInput, SAvgRes* pRes) { + int32_t numOfElems = 0; + float* plist = (float*)pCol->pData; + + if (pCol->hasNull || pInput->numOfRows < 8) { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.dsum += plist[i]; + } + } else { // no null values exist + numOfElems = pInput->numOfRows; + pRes->count += pInput->numOfRows; + + // 1. an software version to speedup the process by using loop unwinding. + + + + // 2. if both the CPU and OS support SSE4.2, let's try the faster version by using SSE4.2 SIMD + + + + // 3. If both the CPU and OS support AVX, let's employ AVX instruction to speedup this loop + // 3.1 find the start position that are aligned to 32bytes address in memory + int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); + int32_t i = 0; + + int32_t bitWidth = 8; + + int32_t remain = (pInput->numOfRows - startElem) % bitWidth; + int32_t rounds = (pInput->numOfRows - startElem) / bitWidth; + const float* p = &plist[startElem]; + + __m256 loadVal; + __m256 sum = _mm256_setzero_ps(); + + for(; i < rounds; ++i) { + loadVal = _mm256_loadu_ps(p); + sum = _mm256_add_ps(sum, loadVal); + p += bitWidth; + } + + // let sum up the final results + const float* q = (const float*)∑ + pRes->sum.dsum += q[0] + q[1] + q[2] + q[3] + q[4] + q[5] + q[6] + q[7]; + + // calculate the front and the reminder items in array list + for(int32_t j = 0; j < startElem; ++j) { + pRes->sum.dsum += plist[j]; + } + + startElem += rounds * bitWidth; + for(int32_t j = 0; j < remain; ++j) { + pRes->sum.dsum += plist[j + startElem]; + } + } + + return numOfElems; +} + +int32_t getAvgInfoSize() { return (int32_t)sizeof(SAvgRes); } + +bool getAvgFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { + pEnv->calcMemSize = sizeof(SAvgRes); + return true; +} + +bool avgFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo) { + if (!functionSetup(pCtx, pResultInfo)) { + return false; + } + + SAvgRes* pRes = GET_ROWCELL_INTERBUF(pResultInfo); + memset(pRes, 0, sizeof(SAvgRes)); + return true; +} + +int32_t avgFunction(SqlFunctionCtx* pCtx) { + int32_t numOfElem = 0; + + SInputColumnInfoData* pInput = &pCtx->input; + SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; + int32_t type = pInput->pData[0]->info.type; + + SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); + pAvgRes->type = type; + + // computing based on the true data block + SColumnInfoData* pCol = pInput->pData[0]; + + int32_t start = pInput->startRowIndex; + int32_t numOfRows = pInput->numOfRows; + + if (IS_NULL_TYPE(type)) { + numOfElem = 0; + goto _avg_over; + } + + if (pInput->colDataAggIsSet) { + numOfElem = numOfRows - pAgg->numOfNull; + ASSERT(numOfElem >= 0); + + pAvgRes->count += numOfElem; + if (IS_SIGNED_NUMERIC_TYPE(type)) { + pAvgRes->sum.isum += pAgg->sum; + } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { + pAvgRes->sum.usum += pAgg->sum; + } else if (IS_FLOAT_TYPE(type)) { + pAvgRes->sum.dsum += GET_DOUBLE_VAL((const char*)&(pAgg->sum)); + } + } else { // computing based on the true data block + switch (type) { + case TSDB_DATA_TYPE_TINYINT: { + int8_t* plist = (int8_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.isum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_SMALLINT: { + int16_t* plist = (int16_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.isum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_INT: { + int32_t* plist = (int32_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.isum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_BIGINT: { + int64_t* plist = (int64_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.isum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_UTINYINT: { + uint8_t* plist = (uint8_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.usum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_USMALLINT: { + uint16_t* plist = (uint16_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.usum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_UINT: { + uint32_t* plist = (uint32_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.usum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_UBIGINT: { + uint64_t* plist = (uint64_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.usum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_FLOAT: { + numOfElem = handleFloatCols(pCol, pInput, pAvgRes); +// float* plist = (float*)pCol->pData; +// // float val = 0; +// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { +// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// numOfElem += 1; +// pAvgRes->count += 1; +// pAvgRes->sum.dsum += plist[i]; +// } + // pAvgRes->sum.dsum = val; + break; + } + + case TSDB_DATA_TYPE_DOUBLE: { + double* plist = (double*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.dsum += plist[i]; + } + break; + } + + default: + break; + } + } + +_avg_over: + // data in the check operation are all null, not output + SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); + return TSDB_CODE_SUCCESS; +} + +static void avgTransferInfo(SAvgRes* pInput, SAvgRes* pOutput) { + pOutput->type = pInput->type; + if (IS_SIGNED_NUMERIC_TYPE(pOutput->type)) { + pOutput->sum.isum += pInput->sum.isum; + } else if (IS_UNSIGNED_NUMERIC_TYPE(pOutput->type)) { + pOutput->sum.usum += pInput->sum.usum; + } else { + pOutput->sum.dsum += pInput->sum.dsum; + } + + pOutput->count += pInput->count; +} + +int32_t avgFunctionMerge(SqlFunctionCtx* pCtx) { + SInputColumnInfoData* pInput = &pCtx->input; + SColumnInfoData* pCol = pInput->pData[0]; + ASSERT(pCol->info.type == TSDB_DATA_TYPE_BINARY); + + SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); + + int32_t start = pInput->startRowIndex; + + for (int32_t i = start; i < start + pInput->numOfRows; ++i) { + char* data = colDataGetData(pCol, i); + SAvgRes* pInputInfo = (SAvgRes*)varDataVal(data); + avgTransferInfo(pInputInfo, pInfo); + } + + SET_VAL(GET_RES_INFO(pCtx), 1, 1); + + return TSDB_CODE_SUCCESS; +} + +int32_t avgInvertFunction(SqlFunctionCtx* pCtx) { + int32_t numOfElem = 0; + + // Only the pre-computing information loaded and actual data does not loaded + SInputColumnInfoData* pInput = &pCtx->input; + int32_t type = pInput->pData[0]->info.type; + + SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); + + // computing based on the true data block + SColumnInfoData* pCol = pInput->pData[0]; + + int32_t start = pInput->startRowIndex; + int32_t numOfRows = pInput->numOfRows; + + switch (type) { + case TSDB_DATA_TYPE_TINYINT: { + LIST_AVG_N(pAvgRes->sum.isum, int8_t); + break; + } + case TSDB_DATA_TYPE_SMALLINT: { + LIST_AVG_N(pAvgRes->sum.isum, int16_t); + break; + } + case TSDB_DATA_TYPE_INT: { + LIST_AVG_N(pAvgRes->sum.isum, int32_t); + break; + } + case TSDB_DATA_TYPE_BIGINT: { + LIST_AVG_N(pAvgRes->sum.isum, int64_t); + break; + } + case TSDB_DATA_TYPE_UTINYINT: { + LIST_AVG_N(pAvgRes->sum.usum, uint8_t); + break; + } + case TSDB_DATA_TYPE_USMALLINT: { + LIST_AVG_N(pAvgRes->sum.usum, uint16_t); + break; + } + case TSDB_DATA_TYPE_UINT: { + LIST_AVG_N(pAvgRes->sum.usum, uint32_t); + break; + } + case TSDB_DATA_TYPE_UBIGINT: { + LIST_AVG_N(pAvgRes->sum.usum, uint64_t); + break; + } + case TSDB_DATA_TYPE_FLOAT: { + LIST_AVG_N(pAvgRes->sum.dsum, float); + break; + } + case TSDB_DATA_TYPE_DOUBLE: { + LIST_AVG_N(pAvgRes->sum.dsum, double); + break; + } + default: + break; + } + + // data in the check operation are all null, not output + SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); + return TSDB_CODE_SUCCESS; +} + +int32_t avgCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx) { + SResultRowEntryInfo* pDResInfo = GET_RES_INFO(pDestCtx); + SAvgRes* pDBuf = GET_ROWCELL_INTERBUF(pDResInfo); + + SResultRowEntryInfo* pSResInfo = GET_RES_INFO(pSourceCtx); + SAvgRes* pSBuf = GET_ROWCELL_INTERBUF(pSResInfo); + int16_t type = pDBuf->type == TSDB_DATA_TYPE_NULL ? pSBuf->type : pDBuf->type; + + if (IS_SIGNED_NUMERIC_TYPE(type)) { + pDBuf->sum.isum += pSBuf->sum.isum; + } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { + pDBuf->sum.usum += pSBuf->sum.usum; + } else { + pDBuf->sum.dsum += pSBuf->sum.dsum; + } + pDBuf->count += pSBuf->count; + + return TSDB_CODE_SUCCESS; +} + +int32_t avgFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { + SInputColumnInfoData* pInput = &pCtx->input; + + SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); + int32_t type = pAvgRes->type; + + if (IS_SIGNED_NUMERIC_TYPE(type)) { + pAvgRes->result = pAvgRes->sum.isum / ((double)pAvgRes->count); + } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { + pAvgRes->result = pAvgRes->sum.usum / ((double)pAvgRes->count); + } else { + pAvgRes->result = pAvgRes->sum.dsum / ((double)pAvgRes->count); + } + + // check for overflow + if (isinf(pAvgRes->result) || isnan(pAvgRes->result)) { + GET_RES_INFO(pCtx)->numOfRes = 0; + } + + return functionFinalize(pCtx, pBlock); +} + +int32_t avgPartialFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { + SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); + SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); + int32_t resultBytes = getAvgInfoSize(); + char* res = taosMemoryCalloc(resultBytes + VARSTR_HEADER_SIZE, sizeof(char)); + + memcpy(varDataVal(res), pInfo, resultBytes); + varDataSetLen(res, resultBytes); + + int32_t slotId = pCtx->pExpr->base.resSchema.slotId; + SColumnInfoData* pCol = taosArrayGet(pBlock->pDataBlock, slotId); + + colDataAppend(pCol, pBlock->info.rows, res, false); + + taosMemoryFree(res); + return pResInfo->numOfRes; +} \ No newline at end of file From ea83ae239e263310e9032e55567a79172d6c80c8 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 10 Nov 2022 18:58:10 +0800 Subject: [PATCH 02/83] refactor: do some internal refactor. --- cmake/cmake.define | 2 +- source/libs/function/inc/builtinsimpl.h | 15 + source/libs/function/src/builtinsimpl.c | 1064 ++++++++--------- .../libs/function/src/detail/tavgfunction.c | 25 +- source/libs/function/src/detail/tminmax.c | 820 +++++++++++++ 5 files changed, 1381 insertions(+), 545 deletions(-) create mode 100644 source/libs/function/src/detail/tminmax.c diff --git a/cmake/cmake.define b/cmake/cmake.define index c7ad766a91..dbd6f30b27 100644 --- a/cmake/cmake.define +++ b/cmake/cmake.define @@ -125,7 +125,7 @@ ELSE () MESSAGE("System processor ID: ${CMAKE_SYSTEM_PROCESSOR}") IF (TD_INTEL_64 OR TD_INTEL_32) - ADD_DEFINITIONS("-msse4.2 -mavx") + ADD_DEFINITIONS("-msse4.2 -mavx -mavx2") IF("${FMA_SUPPORT}" MATCHES "true") MESSAGE(STATUS "turn fma function support on") ADD_DEFINITIONS("-mfma") diff --git a/source/libs/function/inc/builtinsimpl.h b/source/libs/function/inc/builtinsimpl.h index c0b1f62fda..2ec882d1de 100644 --- a/source/libs/function/inc/builtinsimpl.h +++ b/source/libs/function/inc/builtinsimpl.h @@ -32,6 +32,21 @@ typedef struct SSumRes { int16_t type; } SSumRes; +typedef struct SMinmaxResInfo { + bool assign; // assign the first value or not + int64_t v; + STuplePos tuplePos; + + STuplePos nullTuplePos; + bool nullTupleSaved; + int16_t type; +} SMinmaxResInfo; +int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc); + +STuplePos saveTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, const STupleKey* pKey); +int32_t updateTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos); +const char* loadTupleData(SqlFunctionCtx* pCtx, const STuplePos* pPos); + bool functionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo); int32_t functionFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock); int32_t functionFinalizeWithResultBuf(SqlFunctionCtx* pCtx, SSDataBlock* pBlock, char* finalResult); diff --git a/source/libs/function/src/builtinsimpl.c b/source/libs/function/src/builtinsimpl.c index fc9c62c68f..26f9c3ad0b 100644 --- a/source/libs/function/src/builtinsimpl.c +++ b/source/libs/function/src/builtinsimpl.c @@ -48,15 +48,15 @@ // int16_t type; // store the original input type, used in merge function //} SAvgRes; -typedef struct SMinmaxResInfo { - bool assign; // assign the first value or not - int64_t v; - STuplePos tuplePos; - - STuplePos nullTuplePos; - bool nullTupleSaved; - int16_t type; -} SMinmaxResInfo; +//typedef struct SMinmaxResInfo { +// bool assign; // assign the first value or not +// int64_t v; +// STuplePos tuplePos; +// +// STuplePos nullTuplePos; +// bool nullTupleSaved; +// int16_t type; +//} SMinmaxResInfo; typedef struct STopBotResItem { SVariant v; @@ -1124,529 +1124,529 @@ bool getMinmaxFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { return true; } -static STuplePos saveTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, - const STupleKey* pKey); -static int32_t updateTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos); -static const char* loadTupleData(SqlFunctionCtx* pCtx, const STuplePos* pPos); +//static STuplePos saveTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, +// const STupleKey* pKey); +//static int32_t updateTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos); +//static const char* loadTupleData(SqlFunctionCtx* pCtx, const STuplePos* pPos); -static int32_t findRowIndex(int32_t start, int32_t num, SColumnInfoData* pCol, const char* tval) { - // the data is loaded, not only the block SMA value - for (int32_t i = start; i < num + start; ++i) { - char* p = colDataGetData(pCol, i); - if (memcmp((void*)tval, p, pCol->info.bytes) == 0) { - return i; - } - } +//static int32_t findRowIndex(int32_t start, int32_t num, SColumnInfoData* pCol, const char* tval) { +// // the data is loaded, not only the block SMA value +// for (int32_t i = start; i < num + start; ++i) { +// char* p = colDataGetData(pCol, i); +// if (memcmp((void*)tval, p, pCol->info.bytes) == 0) { +// return i; +// } +// } +// +// // if reach here means real data of block SMA is not set in pCtx->input. +// return -1; +//} - // if reach here means real data of block SMA is not set in pCtx->input. - return -1; -} - -int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { - int32_t numOfElems = 0; - - SInputColumnInfoData* pInput = &pCtx->input; - SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; - - SColumnInfoData* pCol = pInput->pData[0]; - int32_t type = pCol->info.type; - - SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); - SMinmaxResInfo* pBuf = GET_ROWCELL_INTERBUF(pResInfo); - pBuf->type = type; - - if (IS_NULL_TYPE(type)) { - numOfElems = 0; - goto _min_max_over; - } - - // data in current data block are qualified to the query - if (pInput->colDataAggIsSet) { - numOfElems = pInput->numOfRows - pAgg->numOfNull; - ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); - if (numOfElems == 0) { - return numOfElems; - } - - void* tval = NULL; - int16_t index = 0; - - if (isMinFunc) { - tval = &pInput->pColumnDataAgg[0]->min; - } else { - tval = &pInput->pColumnDataAgg[0]->max; - } - - if (!pBuf->assign) { - pBuf->v = *(int64_t*)tval; - if (pCtx->subsidiaries.num > 0) { - index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); - if (index >= 0) { - pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); - } - } - } else { - if (IS_SIGNED_NUMERIC_TYPE(type)) { - int64_t prev = 0; - GET_TYPED_DATA(prev, int64_t, type, &pBuf->v); - - int64_t val = GET_INT64_VAL(tval); - if ((prev < val) ^ isMinFunc) { - *(int64_t*)&pBuf->v = val; - if (pCtx->subsidiaries.num > 0) { - index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); - if (index >= 0) { - pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); - } - } - } - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - uint64_t prev = 0; - GET_TYPED_DATA(prev, uint64_t, type, &pBuf->v); - - uint64_t val = GET_UINT64_VAL(tval); - if ((prev < val) ^ isMinFunc) { - *(uint64_t*)&pBuf->v = val; - if (pCtx->subsidiaries.num > 0) { - index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); - if (index >= 0) { - pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); - } - } - } - } else if (type == TSDB_DATA_TYPE_DOUBLE) { - double prev = 0; - GET_TYPED_DATA(prev, double, type, &pBuf->v); - - double val = GET_DOUBLE_VAL(tval); - if ((prev < val) ^ isMinFunc) { - *(double*)&pBuf->v = val; - if (pCtx->subsidiaries.num > 0) { - index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); - if (index >= 0) { - pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); - } - } - } - } else if (type == TSDB_DATA_TYPE_FLOAT) { - float prev = 0; - GET_TYPED_DATA(prev, float, type, &pBuf->v); - - float val = GET_DOUBLE_VAL(tval); - if ((prev < val) ^ isMinFunc) { - *(float*)&pBuf->v = val; - } - - if (pCtx->subsidiaries.num > 0) { - index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); - if (index >= 0) { - pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); - } - } - } - } - - pBuf->assign = true; - return numOfElems; - } - - int32_t start = pInput->startRowIndex; - int32_t numOfRows = pInput->numOfRows; - - if (IS_SIGNED_NUMERIC_TYPE(type) || type == TSDB_DATA_TYPE_BOOL) { - if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) { - int8_t* pData = (int8_t*)pCol->pData; - int8_t* val = (int8_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } else if (type == TSDB_DATA_TYPE_SMALLINT) { - int16_t* pData = (int16_t*)pCol->pData; - int16_t* val = (int16_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } else if (type == TSDB_DATA_TYPE_INT) { - int32_t* pData = (int32_t*)pCol->pData; - int32_t* val = (int32_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } else if (type == TSDB_DATA_TYPE_BIGINT) { - int64_t* pData = (int64_t*)pCol->pData; - int64_t* val = (int64_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - if (type == TSDB_DATA_TYPE_UTINYINT) { - uint8_t* pData = (uint8_t*)pCol->pData; - uint8_t* val = (uint8_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } else if (type == TSDB_DATA_TYPE_USMALLINT) { - uint16_t* pData = (uint16_t*)pCol->pData; - uint16_t* val = (uint16_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } else if (type == TSDB_DATA_TYPE_UINT) { - uint32_t* pData = (uint32_t*)pCol->pData; - uint32_t* val = (uint32_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } else if (type == TSDB_DATA_TYPE_UBIGINT) { - uint64_t* pData = (uint64_t*)pCol->pData; - uint64_t* val = (uint64_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } - } else if (type == TSDB_DATA_TYPE_DOUBLE) { - double* pData = (double*)pCol->pData; - double* val = (double*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } else if (type == TSDB_DATA_TYPE_FLOAT) { - float* pData = (float*)pCol->pData; - float* val = (float*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { -#if 0 - if ((*val) == pData[i]) { - continue; - } - - if ((*val < pData[i]) ^ isMinFunc) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } -#endif - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } - } - -_min_max_over: - if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { - pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); - pBuf->nullTupleSaved = true; - } - return numOfElems; -} +//int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { +// int32_t numOfElems = 0; +// +// SInputColumnInfoData* pInput = &pCtx->input; +// SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; +// +// SColumnInfoData* pCol = pInput->pData[0]; +// int32_t type = pCol->info.type; +// +// SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); +// SMinmaxResInfo* pBuf = GET_ROWCELL_INTERBUF(pResInfo); +// pBuf->type = type; +// +// if (IS_NULL_TYPE(type)) { +// numOfElems = 0; +// goto _min_max_over; +// } +// +// // data in current data block are qualified to the query +// if (pInput->colDataAggIsSet) { +// numOfElems = pInput->numOfRows - pAgg->numOfNull; +// ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); +// if (numOfElems == 0) { +// return numOfElems; +// } +// +// void* tval = NULL; +// int16_t index = 0; +// +// if (isMinFunc) { +// tval = &pInput->pColumnDataAgg[0]->min; +// } else { +// tval = &pInput->pColumnDataAgg[0]->max; +// } +// +// if (!pBuf->assign) { +// pBuf->v = *(int64_t*)tval; +// if (pCtx->subsidiaries.num > 0) { +// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); +// if (index >= 0) { +// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); +// } +// } +// } else { +// if (IS_SIGNED_NUMERIC_TYPE(type)) { +// int64_t prev = 0; +// GET_TYPED_DATA(prev, int64_t, type, &pBuf->v); +// +// int64_t val = GET_INT64_VAL(tval); +// if ((prev < val) ^ isMinFunc) { +// *(int64_t*)&pBuf->v = val; +// if (pCtx->subsidiaries.num > 0) { +// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); +// if (index >= 0) { +// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); +// } +// } +// } +// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { +// uint64_t prev = 0; +// GET_TYPED_DATA(prev, uint64_t, type, &pBuf->v); +// +// uint64_t val = GET_UINT64_VAL(tval); +// if ((prev < val) ^ isMinFunc) { +// *(uint64_t*)&pBuf->v = val; +// if (pCtx->subsidiaries.num > 0) { +// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); +// if (index >= 0) { +// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); +// } +// } +// } +// } else if (type == TSDB_DATA_TYPE_DOUBLE) { +// double prev = 0; +// GET_TYPED_DATA(prev, double, type, &pBuf->v); +// +// double val = GET_DOUBLE_VAL(tval); +// if ((prev < val) ^ isMinFunc) { +// *(double*)&pBuf->v = val; +// if (pCtx->subsidiaries.num > 0) { +// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); +// if (index >= 0) { +// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); +// } +// } +// } +// } else if (type == TSDB_DATA_TYPE_FLOAT) { +// float prev = 0; +// GET_TYPED_DATA(prev, float, type, &pBuf->v); +// +// float val = GET_DOUBLE_VAL(tval); +// if ((prev < val) ^ isMinFunc) { +// *(float*)&pBuf->v = val; +// } +// +// if (pCtx->subsidiaries.num > 0) { +// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); +// if (index >= 0) { +// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); +// } +// } +// } +// } +// +// pBuf->assign = true; +// return numOfElems; +// } +// +// int32_t start = pInput->startRowIndex; +// int32_t numOfRows = pInput->numOfRows; +// +// if (IS_SIGNED_NUMERIC_TYPE(type) || type == TSDB_DATA_TYPE_BOOL) { +// if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) { +// int8_t* pData = (int8_t*)pCol->pData; +// int8_t* val = (int8_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } else if (type == TSDB_DATA_TYPE_SMALLINT) { +// int16_t* pData = (int16_t*)pCol->pData; +// int16_t* val = (int16_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } else if (type == TSDB_DATA_TYPE_INT) { +// int32_t* pData = (int32_t*)pCol->pData; +// int32_t* val = (int32_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } else if (type == TSDB_DATA_TYPE_BIGINT) { +// int64_t* pData = (int64_t*)pCol->pData; +// int64_t* val = (int64_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } +// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { +// if (type == TSDB_DATA_TYPE_UTINYINT) { +// uint8_t* pData = (uint8_t*)pCol->pData; +// uint8_t* val = (uint8_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } else if (type == TSDB_DATA_TYPE_USMALLINT) { +// uint16_t* pData = (uint16_t*)pCol->pData; +// uint16_t* val = (uint16_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } else if (type == TSDB_DATA_TYPE_UINT) { +// uint32_t* pData = (uint32_t*)pCol->pData; +// uint32_t* val = (uint32_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } else if (type == TSDB_DATA_TYPE_UBIGINT) { +// uint64_t* pData = (uint64_t*)pCol->pData; +// uint64_t* val = (uint64_t*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } +// } else if (type == TSDB_DATA_TYPE_DOUBLE) { +// double* pData = (double*)pCol->pData; +// double* val = (double*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +// // ignore the equivalent data value +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } else if (type == TSDB_DATA_TYPE_FLOAT) { +// float* pData = (float*)pCol->pData; +// float* val = (float*)&pBuf->v; +// +// for (int32_t i = start; i < start + numOfRows; ++i) { +// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { +// continue; +// } +// +// if (!pBuf->assign) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); +// } +// pBuf->assign = true; +// } else { +//#if 0 +// if ((*val) == pData[i]) { +// continue; +// } +// +// if ((*val < pData[i]) ^ isMinFunc) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +//#endif +// // NOTE: An faster version to avoid one additional comparison with FPU. +// if (isMinFunc) { // min +// if (*val > pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } else { // max +// if (*val < pData[i]) { +// *val = pData[i]; +// if (pCtx->subsidiaries.num > 0) { +// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); +// } +// } +// } +// } +// +// numOfElems += 1; +// } +// } +// +//_min_max_over: +// if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { +// pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); +// pBuf->nullTupleSaved = true; +// } +// return numOfElems; +//} int32_t minFunction(SqlFunctionCtx* pCtx) { int32_t numOfElems = doMinMaxHelper(pCtx, 1); @@ -3844,7 +3844,7 @@ static int32_t doUpdateTupleData(SSerializeDataHandle* pHandle, const void* pBuf return TSDB_CODE_SUCCESS; } -static int32_t updateTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos) { +int32_t updateTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos) { char* buf = serializeTupleData(pSrcBlock, rowIndex, &pCtx->subsidiaries, pCtx->subsidiaries.buf); doUpdateTupleData(&pCtx->saveHandle, buf, pCtx->subsidiaries.rowLen, pPos); return TSDB_CODE_SUCCESS; @@ -3864,7 +3864,7 @@ static char* doLoadTupleData(SSerializeDataHandle* pHandle, const STuplePos* pPo } } -static const char* loadTupleData(SqlFunctionCtx* pCtx, const STuplePos* pPos) { +const char* loadTupleData(SqlFunctionCtx* pCtx, const STuplePos* pPos) { return doLoadTupleData(&pCtx->saveHandle, pPos); } diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 431e169346..50a69a4241 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -282,19 +282,20 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { } case TSDB_DATA_TYPE_FLOAT: { +#if 1 numOfElem = handleFloatCols(pCol, pInput, pAvgRes); -// float* plist = (float*)pCol->pData; -// // float val = 0; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.dsum += plist[i]; -// } - // pAvgRes->sum.dsum = val; +#else + float* plist = (float*)pCol->pData; + for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { + if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElem += 1; + pAvgRes->count += 1; + pAvgRes->sum.dsum += plist[i]; + } +#endif break; } diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c new file mode 100644 index 0000000000..7814a41f4f --- /dev/null +++ b/source/libs/function/src/detail/tminmax.c @@ -0,0 +1,820 @@ +/* + * Copyright (c) 2019 TAOS Data, Inc. + * + * This program is free software: you can use, redistribute, and/or modify + * it under the terms of the GNU Affero General Public License, version 3 + * or later ("AGPL"), as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. + * + * You should have received a copy of the GNU Affero General Public License + * along with this program. If not, see . + */ + +#include +#include "builtinsimpl.h" +#include "function.h" +#include "tdatablock.h" +#include "tfunctionInt.h" +#include "tglobal.h" + +static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, + SMinmaxResInfo* pBuf, bool isMinFunc) { + int32_t* pData = (int32_t*)pCol->pData; + int32_t* val = (int32_t*)&pBuf->v; + + int32_t numOfElems = 0; + if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + + numOfElems += 1; + } + } else { // max function + for (int32_t i = start; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + + numOfElems += 1; + } + } + } else { // not has null value + // 1. software version + + + + + // 3. AVX2 version to speedup the loop + int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); + int32_t i = 0; + + int32_t bitWidth = 8; + int32_t v = 0; + + int32_t remain = (numOfRows - startElem) % bitWidth; + int32_t rounds = (numOfRows - startElem) / bitWidth; + const int32_t* p = &pData[startElem]; + + __m256i next; + __m256i initialVal = _mm256_loadu_si256((__m256i*)p); + p += bitWidth; + + if (!isMinFunc) { // max function + for (; i < rounds; ++i) { + next = _mm256_loadu_si256((__m256i*)p); + initialVal = _mm256_max_epi32(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const int32_t* q = (const int32_t*)&initialVal; + + v = TMAX(q[0], q[1]); + v = TMAX(v, q[2]); + v = TMAX(v, q[3]); + v = TMAX(v, q[4]); + v = TMAX(v, q[5]); + v = TMAX(v, q[6]); + v = TMAX(v, q[7]); + + // calculate the front and the reminder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + startElem]) { + v = p[j + startElem]; + } + } + } else { // min function + for (; i < rounds; ++i) { + next = _mm256_loadu_si256((__m256i*)p); + initialVal = _mm256_min_epi32(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const int32_t* q = (const int32_t*)&initialVal; + + v = TMIN(q[0], q[1]); + v = TMIN(v, q[2]); + v = TMIN(v, q[3]); + v = TMIN(v, q[4]); + v = TMIN(v, q[5]); + v = TMIN(v, q[6]); + v = TMIN(v, q[7]); + + // calculate the front and the reminder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + startElem]) { + v = p[j + startElem]; + } + } + } + + *val = v; + numOfElems = numOfRows; + } + + return numOfElems; +} + +static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, + SMinmaxResInfo* pBuf, bool isMinFunc) { + float* pData = (float*)pCol->pData; + double* val = (double*)&pBuf->v; + + int32_t numOfElems = 0; + if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + + numOfElems += 1; + } + } else { // max function + for (int32_t i = start; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + + numOfElems += 1; + } + } + } else { // not has null value + // 1. software version + + + + + // 3. AVX2 version to speedup the loop + int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); + int32_t i = 0; + + int32_t bitWidth = 8; + float v = 0; + + int32_t remain = (numOfRows - startElem) % bitWidth; + int32_t rounds = (numOfRows - startElem) / bitWidth; + const float* p = &pData[startElem]; + + __m256 next; + __m256 initialVal = _mm256_loadu_ps(p); + p += bitWidth; + + if (!isMinFunc) { // max function + for (; i < rounds; ++i) { + next = _mm256_loadu_ps(p); + initialVal = _mm256_max_ps(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const float* q = (const float*)&initialVal; + + v = TMAX(q[0], q[1]); + v = TMAX(v, q[2]); + v = TMAX(v, q[3]); + v = TMAX(v, q[4]); + v = TMAX(v, q[5]); + v = TMAX(v, q[6]); + v = TMAX(v, q[7]); + + // calculate the front and the reminder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + startElem]) { + v = p[j + startElem]; + } + } + } else { // min function + for (; i < rounds; ++i) { + next = _mm256_loadu_ps(p); + initialVal = _mm256_min_ps(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const float* q = (const float*)&initialVal; + + v = TMIN(q[0], q[1]); + v = TMIN(v, q[2]); + v = TMIN(v, q[3]); + v = TMIN(v, q[4]); + v = TMIN(v, q[5]); + v = TMIN(v, q[6]); + v = TMIN(v, q[7]); + + // calculate the front and the reminder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + startElem]) { + v = p[j + startElem]; + } + } + } + + *val = v; + numOfElems = numOfRows; + } + + return numOfElems; +} + +static int32_t findRowIndex(int32_t start, int32_t num, SColumnInfoData* pCol, const char* tval) { + // the data is loaded, not only the block SMA value + for (int32_t i = start; i < num + start; ++i) { + char* p = colDataGetData(pCol, i); + if (memcmp((void*)tval, p, pCol->info.bytes) == 0) { + return i; + } + } + + // if reach here means real data of block SMA is not set in pCtx->input. + return -1; +} + +int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { + int32_t numOfElems = 0; + + SInputColumnInfoData* pInput = &pCtx->input; + SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; + + SColumnInfoData* pCol = pInput->pData[0]; + int32_t type = pCol->info.type; + + SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); + SMinmaxResInfo* pBuf = GET_ROWCELL_INTERBUF(pResInfo); + pBuf->type = type; + + if (IS_NULL_TYPE(type)) { + numOfElems = 0; + goto _min_max_over; + } + + // data in current data block are qualified to the query + if (pInput->colDataAggIsSet) { + numOfElems = pInput->numOfRows - pAgg->numOfNull; + ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); + if (numOfElems == 0) { + return numOfElems; + } + + void* tval = NULL; + int16_t index = 0; + + if (isMinFunc) { + tval = &pInput->pColumnDataAgg[0]->min; + } else { + tval = &pInput->pColumnDataAgg[0]->max; + } + + if (!pBuf->assign) { + pBuf->v = *(int64_t*)tval; + if (pCtx->subsidiaries.num > 0) { + index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); + if (index >= 0) { + pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); + } + } + } else { + if (IS_SIGNED_NUMERIC_TYPE(type)) { + int64_t prev = 0; + GET_TYPED_DATA(prev, int64_t, type, &pBuf->v); + + int64_t val = GET_INT64_VAL(tval); + if ((prev < val) ^ isMinFunc) { + *(int64_t*)&pBuf->v = val; + if (pCtx->subsidiaries.num > 0) { + index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); + if (index >= 0) { + pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); + } + } + } + } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { + uint64_t prev = 0; + GET_TYPED_DATA(prev, uint64_t, type, &pBuf->v); + + uint64_t val = GET_UINT64_VAL(tval); + if ((prev < val) ^ isMinFunc) { + *(uint64_t*)&pBuf->v = val; + if (pCtx->subsidiaries.num > 0) { + index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); + if (index >= 0) { + pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); + } + } + } + } else if (type == TSDB_DATA_TYPE_DOUBLE) { + double prev = 0; + GET_TYPED_DATA(prev, double, type, &pBuf->v); + + double val = GET_DOUBLE_VAL(tval); + if ((prev < val) ^ isMinFunc) { + *(double*)&pBuf->v = val; + if (pCtx->subsidiaries.num > 0) { + index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); + if (index >= 0) { + pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); + } + } + } + } else if (type == TSDB_DATA_TYPE_FLOAT) { + float prev = 0; + GET_TYPED_DATA(prev, float, type, &pBuf->v); + + float val = GET_DOUBLE_VAL(tval); + if ((prev < val) ^ isMinFunc) { + *(float*)&pBuf->v = val; + } + + if (pCtx->subsidiaries.num > 0) { + index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); + if (index >= 0) { + pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); + } + } + } + } + + pBuf->assign = true; + return numOfElems; + } + + int32_t start = pInput->startRowIndex; + int32_t numOfRows = pInput->numOfRows; + + if (IS_SIGNED_NUMERIC_TYPE(type) || type == TSDB_DATA_TYPE_BOOL) { + if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) { + int8_t* pData = (int8_t*)pCol->pData; + int8_t* val = (int8_t*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } else if (type == TSDB_DATA_TYPE_SMALLINT) { + int16_t* pData = (int16_t*)pCol->pData; + int16_t* val = (int16_t*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } else if (type == TSDB_DATA_TYPE_INT) { + int32_t* pData = (int32_t*)pCol->pData; + int32_t* val = (int32_t*)&pBuf->v; + + numOfElems = handleInt32Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); +#if 0 + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } +#endif + + } else if (type == TSDB_DATA_TYPE_BIGINT) { + int64_t* pData = (int64_t*)pCol->pData; + int64_t* val = (int64_t*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } + } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { + if (type == TSDB_DATA_TYPE_UTINYINT) { + uint8_t* pData = (uint8_t*)pCol->pData; + uint8_t* val = (uint8_t*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } else if (type == TSDB_DATA_TYPE_USMALLINT) { + uint16_t* pData = (uint16_t*)pCol->pData; + uint16_t* val = (uint16_t*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } else if (type == TSDB_DATA_TYPE_UINT) { + uint32_t* pData = (uint32_t*)pCol->pData; + uint32_t* val = (uint32_t*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } else if (type == TSDB_DATA_TYPE_UBIGINT) { + uint64_t* pData = (uint64_t*)pCol->pData; + uint64_t* val = (uint64_t*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } + } else if (type == TSDB_DATA_TYPE_DOUBLE) { + double* pData = (double*)pCol->pData; + double* val = (double*)&pBuf->v; + + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } + } else if (type == TSDB_DATA_TYPE_FLOAT) { + float* pData = (float*)pCol->pData; + float* val = (float*)&pBuf->v; + + numOfElems = handleFloatCol(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); +#if 0 + for (int32_t i = start; i < start + numOfRows; ++i) { + if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (!pBuf->assign) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + } else { +#if 0 + if ((*val) == pData[i]) { + continue; + } + + if ((*val < pData[i]) ^ isMinFunc) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } +#endif + // NOTE: An faster version to avoid one additional comparison with FPU. + if (isMinFunc) { // min + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } else { // max + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + } + } + + numOfElems += 1; + } +#endif + + } + +_min_max_over: + if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { + pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); + pBuf->nullTupleSaved = true; + } + return numOfElems; +} \ No newline at end of file From b83f89572655c5c7229231146aeda060c31a5821 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 11 Nov 2022 14:16:13 +0800 Subject: [PATCH 03/83] refactor: do some internal refactor. --- cmake/cmake.define | 14 +- cmake/cmake.platform | 18 +- cmake/cmake.version | 8 +- include/os/os.h | 7 + include/os/osDef.h | 32 +- include/os/osEnv.h | 5 + include/os/osSysinfo.h | 1 + source/common/src/tglobal.c | 10 +- .../libs/function/src/detail/tavgfunction.c | 87 ++--- source/libs/function/src/detail/tminmax.c | 313 ++++++++++-------- source/os/src/osEnv.c | 8 +- source/os/src/osFile.c | 2 + source/os/src/osLocale.c | 5 +- source/os/src/osSysinfo.c | 71 +++- source/os/src/osTime.c | 5 +- source/util/src/tconfig.c | 8 +- source/util/src/tcrc32c.c | 1 - 17 files changed, 359 insertions(+), 236 deletions(-) diff --git a/cmake/cmake.define b/cmake/cmake.define index dbd6f30b27..3b6024efc8 100644 --- a/cmake/cmake.define +++ b/cmake/cmake.define @@ -123,14 +123,20 @@ ELSE () SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Werror -Wno-literal-suffix -Werror=return-type -fPIC -gdwarf-2 -g3 -Wformat=2 -Wno-format-nonliteral -Wno-format-truncation -Wno-format-y2k") ENDIF () - MESSAGE("System processor ID: ${CMAKE_SYSTEM_PROCESSOR}") IF (TD_INTEL_64 OR TD_INTEL_32) - ADD_DEFINITIONS("-msse4.2 -mavx -mavx2") + ADD_DEFINITIONS("-msse4.2") IF("${FMA_SUPPORT}" MATCHES "true") - MESSAGE(STATUS "turn fma function support on") + MESSAGE(STATUS "fma function supported") ADD_DEFINITIONS("-mfma") ELSE () - MESSAGE(STATUS "turn fma function support off") + MESSAGE(STATUS "fma function NOT supported") + ENDIF() + + IF("${SIMD_SUPPORT}" MATCHES "true") + ADD_DEFINITIONS("-mavx -mavx2") + MESSAGE(STATUS "cpu simd instruction AVX/AVX2 supported") + ELSE() + MESSAGE(STATUS "cpu simd instruction AVX/AVX2 NOT supported") ENDIF() ENDIF () diff --git a/cmake/cmake.platform b/cmake/cmake.platform index 3e239d2e0c..c3680e0de4 100644 --- a/cmake/cmake.platform +++ b/cmake/cmake.platform @@ -1,20 +1,17 @@ cmake_minimum_required(VERSION 3.0) -MESSAGE("Current system is ${CMAKE_SYSTEM_NAME}") - # init SET(TD_LINUX FALSE) SET(TD_WINDOWS FALSE) SET(TD_DARWIN FALSE) -MESSAGE("Compiler ID: ${CMAKE_CXX_COMPILER_ID}") if(CMAKE_COMPILER_IS_GNUCXX MATCHES 1) set(CXX_COMPILER_IS_GNU TRUE) else() set(CXX_COMPILER_IS_GNU FALSE) endif() -MESSAGE("Current system name is ${CMAKE_SYSTEM_NAME}.") +MESSAGE("Current system: ${CMAKE_SYSTEM_NAME}") IF (${CMAKE_SYSTEM_NAME} MATCHES "Linux" OR ${CMAKE_SYSTEM_NAME} MATCHES "Darwin") @@ -26,6 +23,8 @@ IF (${CMAKE_SYSTEM_NAME} MATCHES "Linux" OR ${CMAKE_SYSTEM_NAME} MATCHES "Darwin set(CMAKE_SHARED_LIBRARY_CREATE_CXX_FLAGS "${CMAKE_SHARED_LIBRARY_CREATE_CXX_FLAGS} -undefined dynamic_lookup") ENDIF () + MESSAGE("Current system processor: ${CMAKE_SYSTEM_PROCESSOR}") + IF (${CMAKE_SYSTEM_NAME} MATCHES "Linux") SET(TD_LINUX TRUE) @@ -44,7 +43,6 @@ IF (${CMAKE_SYSTEM_NAME} MATCHES "Linux" OR ${CMAKE_SYSTEM_NAME} MATCHES "Darwin SET(OSTYPE "macOS") ADD_DEFINITIONS("-DDARWIN -Wno-tautological-pointer-compare") - MESSAGE("Current system processor is ${CMAKE_SYSTEM_PROCESSOR}.") IF (${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm64") MESSAGE("Current system arch is arm64") SET(TD_DARWIN_64 TRUE) @@ -80,24 +78,22 @@ ELSEIF (${CMAKE_SYSTEM_NAME} MATCHES "Windows") ENDIF() IF ("${CPUTYPE}" STREQUAL "") - MESSAGE(STATUS "The current platform " ${CMAKE_SYSTEM_PROCESSOR} " is detected") - IF (CMAKE_SYSTEM_PROCESSOR MATCHES "(amd64)|(AMD64)") - MESSAGE(STATUS "The current platform is amd64") + MESSAGE(STATUS "Current platform is amd64") SET(PLATFORM_ARCH_STR "amd64") SET(TD_INTEL_64 TRUE) ELSEIF (CMAKE_SYSTEM_PROCESSOR MATCHES "(x86)|(X86)") - MESSAGE(STATUS "The current platform is x86") + MESSAGE(STATUS "Current platform is x86") SET(PLATFORM_ARCH_STR "i386") SET(TD_INTEL_32 TRUE) ELSEIF (CMAKE_SYSTEM_PROCESSOR MATCHES "armv7l") - MESSAGE(STATUS "The current platform is aarch32") + MESSAGE(STATUS "Current platform is aarch32") SET(PLATFORM_ARCH_STR "arm") SET(TD_ARM_32 TRUE) ADD_DEFINITIONS("-D_TD_ARM_") ADD_DEFINITIONS("-D_TD_ARM_32") ELSEIF (CMAKE_SYSTEM_PROCESSOR MATCHES "(aarch64)|(arm64)") - MESSAGE(STATUS "The current platform is aarch64") + MESSAGE(STATUS "Current platform is aarch64") SET(PLATFORM_ARCH_STR "arm64") SET(TD_ARM_64 TRUE) ADD_DEFINITIONS("-D_TD_ARM_") diff --git a/cmake/cmake.version b/cmake/cmake.version index 03598519ed..0447f284f1 100644 --- a/cmake/cmake.version +++ b/cmake/cmake.version @@ -26,7 +26,7 @@ ELSEIF (HAVE_GIT) SET(TD_VER_GIT "no git commit id") ENDIF () ELSE () - message(STATUS "no git cmd") + message(STATUS "no git found") SET(TD_VER_GIT "no git commit id") ENDIF () @@ -70,9 +70,9 @@ MESSAGE(STATUS "compatible: " ${TD_VER_COMPATIBLE}) MESSAGE(STATUS "commit id: " ${TD_VER_GIT}) MESSAGE(STATUS "build date: " ${TD_VER_DATE}) MESSAGE(STATUS "build type: " ${CMAKE_BUILD_TYPE}) -MESSAGE(STATUS "type: " ${TD_VER_VERTYPE}) -MESSAGE(STATUS "cpu: " ${TD_VER_CPUTYPE}) -MESSAGE(STATUS "os: " ${TD_VER_OSTYPE}) +MESSAGE(STATUS "type: " ${TD_VER_VERTYPE}) +MESSAGE(STATUS "cpu: " ${TD_VER_CPUTYPE}) +MESSAGE(STATUS "os: " ${TD_VER_OSTYPE}) MESSAGE(STATUS "============= compile version parameter information end ============= ") STRING(REPLACE "." "_" TD_LIB_VER_NUMBER ${TD_VER_NUMBER}) diff --git a/include/os/os.h b/include/os/os.h index e780611c41..0334cd4d95 100644 --- a/include/os/os.h +++ b/include/os/os.h @@ -81,6 +81,13 @@ extern "C" { #include #include #include +#include + +#if __AVX__ +#include +#elif __SSE4_2__ +#include +#endif #include "osThread.h" diff --git a/include/os/osDef.h b/include/os/osDef.h index 297d19e21a..0bf9c6184e 100644 --- a/include/os/osDef.h +++ b/include/os/osDef.h @@ -168,22 +168,22 @@ void syslog(int unused, const char *format, ...); } \ } while (0) -#define DEFAULT_DOUBLE_COMP(x, y) \ - do { \ - if (isnan(x) && isnan(y)) { \ - return 0; \ - } \ - if (isnan(x)) { \ - return -1; \ - } \ - if (isnan(y)) { \ - return 1; \ - } \ - if ((x) == (y)) { \ - return 0; \ - } else { \ - return (x) < (y) ? -1 : 1; \ - } \ +#define DEFAULT_DOUBLE_COMP(x, y) \ + do { \ + if (isnan(x) && isnan(y)) { \ + return 0; \ + } \ + if (isnan(x)) { \ + return -1; \ + } \ + if (isnan(y)) { \ + return 1; \ + } \ + if (fabs((x) - (y)) <= DBL_EPSILON) { \ + return 0; \ + } else { \ + return (x) < (y) ? -1 : 1; \ + } \ } while (0) #define DEFAULT_FLOAT_COMP(x, y) DEFAULT_DOUBLE_COMP(x, y) diff --git a/include/os/osEnv.h b/include/os/osEnv.h index c1fdc9e404..a3bd209693 100644 --- a/include/os/osEnv.h +++ b/include/os/osEnv.h @@ -36,6 +36,11 @@ extern int64_t tsStreamMax; extern float tsNumOfCores; extern int64_t tsTotalMemoryKB; extern char *tsProcPath; +extern char tsSIMDEnable; +extern char tsSSE42Enable; +extern char tsAVXEnable; +extern char tsAVX2Enable; +extern char tsFMAEnable; extern char configDir[]; extern char tsDataDir[]; diff --git a/include/os/osSysinfo.h b/include/os/osSysinfo.h index 47cdb02a6f..7765a60f88 100644 --- a/include/os/osSysinfo.h +++ b/include/os/osSysinfo.h @@ -40,6 +40,7 @@ int32_t taosGetOsReleaseName(char *releaseName, int32_t maxLen); int32_t taosGetCpuInfo(char *cpuModel, int32_t maxLen, float *numOfCores); int32_t taosGetCpuCores(float *numOfCores); void taosGetCpuUsage(double *cpu_system, double *cpu_engine); +int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma); int32_t taosGetTotalMemory(int64_t *totalKB); int32_t taosGetProcMemory(int64_t *usedKB); int32_t taosGetSysMemory(int64_t *usedKB); diff --git a/source/common/src/tglobal.c b/source/common/src/tglobal.c index 1be77077b6..50b2c976fd 100644 --- a/source/common/src/tglobal.c +++ b/source/common/src/tglobal.c @@ -15,7 +15,6 @@ #define _DEFAULT_SOURCE #include "tglobal.h" -#include "tcompare.h" #include "tconfig.h" #include "tdatablock.h" #include "tgrant.h" @@ -312,7 +311,14 @@ static int32_t taosAddSystemCfg(SConfig *pCfg) { if (cfgAddLocale(pCfg, "locale", tsLocale) != 0) return -1; if (cfgAddCharset(pCfg, "charset", tsCharset) != 0) return -1; if (cfgAddBool(pCfg, "enableCoreFile", 1, 1) != 0) return -1; - if (cfgAddFloat(pCfg, "numOfCores", tsNumOfCores, 0, 100000, 1) != 0) return -1; + if (cfgAddFloat(pCfg, "numOfCores", tsNumOfCores, 1, 100000, 1) != 0) return -1; + + if (cfgAddBool(pCfg, "SSE42", tsSSE42Enable, 0) != 0) return -1; + if (cfgAddBool(pCfg, "AVX", tsAVXEnable, 0) != 0) return -1; + if (cfgAddBool(pCfg, "AVX2", tsAVX2Enable, 0) != 0) return -1; + if (cfgAddBool(pCfg, "FMA", tsFMAEnable, 0) != 0) return -1; + if (cfgAddBool(pCfg, "SIMD-Supported", tsSIMDEnable, 0) != 0) return -1; + if (cfgAddInt64(pCfg, "openMax", tsOpenMax, 0, INT64_MAX, 1) != 0) return -1; if (cfgAddInt64(pCfg, "streamMax", tsStreamMax, 0, INT64_MAX, 1) != 0) return -1; if (cfgAddInt32(pCfg, "pageSizeKB", tsPageSizeKB, 0, INT64_MAX, 1) != 0) return -1; diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 50a69a4241..01e0a499eb 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -13,7 +13,6 @@ * along with this program. If not, see . */ -#include #include "builtinsimpl.h" #include "function.h" #include "tdatablock.h" @@ -49,11 +48,48 @@ typedef struct SAvgRes { int16_t type; // store the original input type, used in merge function } SAvgRes; +static void floatVectorSumAVX(const SInputColumnInfoData* pInput, const float* plist, SAvgRes* pRes) { +#if __AVX__ + // find the start position that are aligned to 32bytes address in memory + int32_t startIndex = 0; //((uint64_t)plist) & ((1<<8u)-1); + int32_t bitWidth = 8; + + int32_t remain = (pInput->numOfRows - startIndex) % bitWidth; + int32_t rounds = (pInput->numOfRows - startIndex) / bitWidth; + const float* p = &plist[startIndex]; + + __m256 val; + __m256 sum = _mm256_setzero_ps(); + + for (int32_t i = 0; i < rounds; ++i) { + val = _mm256_loadu_ps(p); + sum = _mm256_add_ps(sum, val); + p += bitWidth; + } + + // let sum up the final results + const float* q = (const float*)∑ + pRes->sum.dsum += q[0] + q[1] + q[2] + q[3] + q[4] + q[5] + q[6] + q[7]; + + // calculate the front and the reminder items in array list + for (int32_t j = 0; j < startIndex; ++j) { + pRes->sum.dsum += plist[j]; + } + + startIndex += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + pRes->sum.dsum += plist[j + startIndex]; + } +#endif +} + static int32_t handleFloatCols(const SColumnInfoData* pCol, const SInputColumnInfoData* pInput, SAvgRes* pRes) { int32_t numOfElems = 0; float* plist = (float*)pCol->pData; - if (pCol->hasNull || pInput->numOfRows < 8) { + const int32_t THRESHOLD_SIZE = 8; + + if (pCol->hasNull || pInput->numOfRows <= THRESHOLD_SIZE) { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { if (colDataIsNull_f(pCol->nullbitmap, i)) { continue; @@ -67,46 +103,13 @@ static int32_t handleFloatCols(const SColumnInfoData* pCol, const SInputColumnIn numOfElems = pInput->numOfRows; pRes->count += pInput->numOfRows; - // 1. an software version to speedup the process by using loop unwinding. - - - - // 2. if both the CPU and OS support SSE4.2, let's try the faster version by using SSE4.2 SIMD - - - - // 3. If both the CPU and OS support AVX, let's employ AVX instruction to speedup this loop - // 3.1 find the start position that are aligned to 32bytes address in memory - int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); - int32_t i = 0; - - int32_t bitWidth = 8; - - int32_t remain = (pInput->numOfRows - startElem) % bitWidth; - int32_t rounds = (pInput->numOfRows - startElem) / bitWidth; - const float* p = &plist[startElem]; - - __m256 loadVal; - __m256 sum = _mm256_setzero_ps(); - - for(; i < rounds; ++i) { - loadVal = _mm256_loadu_ps(p); - sum = _mm256_add_ps(sum, loadVal); - p += bitWidth; - } - - // let sum up the final results - const float* q = (const float*)∑ - pRes->sum.dsum += q[0] + q[1] + q[2] + q[3] + q[4] + q[5] + q[6] + q[7]; - - // calculate the front and the reminder items in array list - for(int32_t j = 0; j < startElem; ++j) { - pRes->sum.dsum += plist[j]; - } - - startElem += rounds * bitWidth; - for(int32_t j = 0; j < remain; ++j) { - pRes->sum.dsum += plist[j + startElem]; + // 3. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (tsAVXEnable && tsSIMDEnable) { + floatVectorSumAVX(pInput, plist, pRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pRes->sum.dsum += plist[i]; + } } } diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index 7814a41f4f..074e5ef428 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -13,20 +13,163 @@ * along with this program. If not, see . */ -#include #include "builtinsimpl.h" #include "function.h" #include "tdatablock.h" #include "tfunctionInt.h" #include "tglobal.h" +static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool isMinFunc) { + int32_t v = 0; + +#if __AVX2__ + int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); + int32_t bitWidth = 8; + + int32_t remain = (numOfRows - startElem) % bitWidth; + int32_t rounds = (numOfRows - startElem) / bitWidth; + const int32_t* p = &pData[startElem]; + + __m256i next; + __m256i initialVal = _mm256_loadu_si256((__m256i*)p); + p += bitWidth; + + if (!isMinFunc) { // max function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_loadu_si256((__m256i*)p); + initialVal = _mm256_max_epi32(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const int32_t* q = (const int32_t*)&initialVal; + + v = TMAX(q[0], q[1]); + v = TMAX(v, q[2]); + v = TMAX(v, q[3]); + v = TMAX(v, q[4]); + v = TMAX(v, q[5]); + v = TMAX(v, q[6]); + v = TMAX(v, q[7]); + + // calculate the front and the reminder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + startElem]) { + v = p[j + startElem]; + } + } + } else { // min function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_loadu_si256((__m256i*)p); + initialVal = _mm256_min_epi32(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const int32_t* q = (const int32_t*)&initialVal; + + v = TMIN(q[0], q[1]); + v = TMIN(v, q[2]); + v = TMIN(v, q[3]); + v = TMIN(v, q[4]); + v = TMIN(v, q[5]); + v = TMIN(v, q[6]); + v = TMIN(v, q[7]); + + // calculate the front and the remainder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + startElem]) { + v = p[j + startElem]; + } + } + } +#endif + + return v; +} + +static float floatVectorCmpAVX(const float* pData, int32_t numOfRows, bool isMinFunc) { + float v = 0; + +#if __AVX__ + int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); + int32_t i = 0; + + int32_t bitWidth = 8; + + int32_t remain = (numOfRows - startElem) % bitWidth; + int32_t rounds = (numOfRows - startElem) / bitWidth; + const float* p = &pData[startElem]; + + __m256 next; + __m256 initialVal = _mm256_loadu_ps(p); + p += bitWidth; + + if (!isMinFunc) { // max function + for (; i < rounds; ++i) { + next = _mm256_loadu_ps(p); + initialVal = _mm256_max_ps(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const float* q = (const float*)&initialVal; + + v = TMAX(q[0], q[1]); + v = TMAX(v, q[2]); + v = TMAX(v, q[3]); + v = TMAX(v, q[4]); + v = TMAX(v, q[5]); + v = TMAX(v, q[6]); + v = TMAX(v, q[7]); + + // calculate the front and the reminder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + startElem]) { + v = p[j + startElem]; + } + } + } else { // min function + for (; i < rounds; ++i) { + next = _mm256_loadu_ps(p); + initialVal = _mm256_min_ps(initialVal, next); + p += bitWidth; + } + + // let sum up the final results + const float* q = (const float*)&initialVal; + + v = TMIN(q[0], q[1]); + v = TMIN(v, q[2]); + v = TMIN(v, q[3]); + v = TMIN(v, q[4]); + v = TMIN(v, q[5]); + v = TMIN(v, q[6]); + v = TMIN(v, q[7]); + + // calculate the front and the reminder items in array list + startElem += rounds * bitWidth; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + startElem]) { + v = p[j + startElem]; + } + } + } +#endif + + return v; +} + static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { int32_t* pData = (int32_t*)pCol->pData; int32_t* val = (int32_t*)&pBuf->v; int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { + if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { if (isMinFunc) { // min for (int32_t i = start; i < start + numOfRows; ++i) { if (colDataIsNull_f(pCol->nullbitmap, i)) { @@ -77,79 +220,30 @@ static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numO } } } else { // not has null value - // 1. software version - - - - - // 3. AVX2 version to speedup the loop - int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); - int32_t i = 0; - - int32_t bitWidth = 8; - int32_t v = 0; - - int32_t remain = (numOfRows - startElem) % bitWidth; - int32_t rounds = (numOfRows - startElem) / bitWidth; - const int32_t* p = &pData[startElem]; - - __m256i next; - __m256i initialVal = _mm256_loadu_si256((__m256i*)p); - p += bitWidth; - - if (!isMinFunc) { // max function - for (; i < rounds; ++i) { - next = _mm256_loadu_si256((__m256i*)p); - initialVal = _mm256_max_epi32(initialVal, next); - p += bitWidth; + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + *val = i32VectorCmpAVX2(pData, numOfRows, isMinFunc); + } else { + if (!pBuf->assign) { + *val = pData[0]; + pBuf->assign = true; } - // let sum up the final results - const int32_t* q = (const int32_t*)&initialVal; - - v = TMAX(q[0], q[1]); - v = TMAX(v, q[2]); - v = TMAX(v, q[3]); - v = TMAX(v, q[4]); - v = TMAX(v, q[5]); - v = TMAX(v, q[6]); - v = TMAX(v, q[7]); - - // calculate the front and the reminder items in array list - startElem += rounds * bitWidth; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + startElem]) { - v = p[j + startElem]; + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val > pData[i]) { + *val = pData[i]; + } } - } - } else { // min function - for (; i < rounds; ++i) { - next = _mm256_loadu_si256((__m256i*)p); - initialVal = _mm256_min_epi32(initialVal, next); - p += bitWidth; - } - - // let sum up the final results - const int32_t* q = (const int32_t*)&initialVal; - - v = TMIN(q[0], q[1]); - v = TMIN(v, q[2]); - v = TMIN(v, q[3]); - v = TMIN(v, q[4]); - v = TMIN(v, q[5]); - v = TMIN(v, q[6]); - v = TMIN(v, q[7]); - - // calculate the front and the reminder items in array list - startElem += rounds * bitWidth; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + startElem]) { - v = p[j + startElem]; + } else { // max + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val < pData[i]) { + *val = pData[i]; + } } } } - *val = v; numOfElems = numOfRows; } @@ -213,79 +307,30 @@ static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numO } } } else { // not has null value - // 1. software version - - - - - // 3. AVX2 version to speedup the loop - int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); - int32_t i = 0; - - int32_t bitWidth = 8; - float v = 0; - - int32_t remain = (numOfRows - startElem) % bitWidth; - int32_t rounds = (numOfRows - startElem) / bitWidth; - const float* p = &pData[startElem]; - - __m256 next; - __m256 initialVal = _mm256_loadu_ps(p); - p += bitWidth; - - if (!isMinFunc) { // max function - for (; i < rounds; ++i) { - next = _mm256_loadu_ps(p); - initialVal = _mm256_max_ps(initialVal, next); - p += bitWidth; + // AVX version to speedup the loop + if (tsAVXEnable && tsSIMDEnable) { + *val = (double) floatVectorCmpAVX(pData, numOfRows, isMinFunc); + } else { + if (!pBuf->assign) { + *val = pData[0]; + pBuf->assign = true; } - // let sum up the final results - const float* q = (const float*)&initialVal; - - v = TMAX(q[0], q[1]); - v = TMAX(v, q[2]); - v = TMAX(v, q[3]); - v = TMAX(v, q[4]); - v = TMAX(v, q[5]); - v = TMAX(v, q[6]); - v = TMAX(v, q[7]); - - // calculate the front and the reminder items in array list - startElem += rounds * bitWidth; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + startElem]) { - v = p[j + startElem]; + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val > pData[i]) { + *val = pData[i]; + } } - } - } else { // min function - for (; i < rounds; ++i) { - next = _mm256_loadu_ps(p); - initialVal = _mm256_min_ps(initialVal, next); - p += bitWidth; - } - - // let sum up the final results - const float* q = (const float*)&initialVal; - - v = TMIN(q[0], q[1]); - v = TMIN(v, q[2]); - v = TMIN(v, q[3]); - v = TMIN(v, q[4]); - v = TMIN(v, q[5]); - v = TMIN(v, q[6]); - v = TMIN(v, q[7]); - - // calculate the front and the reminder items in array list - startElem += rounds * bitWidth; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + startElem]) { - v = p[j + startElem]; + } else { // max + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val < pData[i]) { + *val = pData[i]; + } } } } - *val = v; numOfElems = numOfRows; } diff --git a/source/os/src/osEnv.c b/source/os/src/osEnv.c index ac1881fc6d..7063d1f574 100644 --- a/source/os/src/osEnv.c +++ b/source/os/src/osEnv.c @@ -37,6 +37,12 @@ float tsNumOfCores = 0; int64_t tsTotalMemoryKB = 0; char *tsProcPath = NULL; +char tsSIMDEnable = 0; +char tsSSE42Enable = 0; +char tsAVXEnable = 0; +char tsAVX2Enable = 0; +char tsFMAEnable = 0; + void osDefaultInit() { taosSeedRand(taosSafeRand()); taosGetSystemLocale(tsLocale, tsCharset); @@ -99,7 +105,7 @@ bool osDataSpaceSufficient() { return tsDataSpace.size.avail > tsDataSpace.reser bool osTempSpaceSufficient() { return tsTempSpace.size.avail > tsTempSpace.reserved; } -void osSetTimezone(const char *timezone) { taosSetSystemTimezone(timezone, tsTimezoneStr, &tsDaylight, &tsTimezone); } +void osSetTimezone(const char *tz) { taosSetSystemTimezone(tz, tsTimezoneStr, &tsDaylight, &tsTimezone); } void osSetSystemLocale(const char *inLocale, const char *inCharSet) { memcpy(tsLocale, inLocale, strlen(inLocale) + 1); diff --git a/source/os/src/osFile.c b/source/os/src/osFile.c index 94a10322ed..9b42a7ea44 100644 --- a/source/os/src/osFile.c +++ b/source/os/src/osFile.c @@ -775,6 +775,7 @@ int64_t taosGetLineFile(TdFilePtr pFile, char **__restrict ptrBuf) { return getline(ptrBuf, &len, pFile->fp); #endif } + int64_t taosGetsFile(TdFilePtr pFile, int32_t maxSize, char *__restrict buf) { if (pFile == NULL || buf == NULL) { return -1; @@ -785,6 +786,7 @@ int64_t taosGetsFile(TdFilePtr pFile, int32_t maxSize, char *__restrict buf) { } return strlen(buf); } + int32_t taosEOFFile(TdFilePtr pFile) { if (pFile == NULL) { return 0; diff --git a/source/os/src/osLocale.c b/source/os/src/osLocale.c index 89216ecaf4..7319181a77 100644 --- a/source/os/src/osLocale.c +++ b/source/os/src/osLocale.c @@ -67,6 +67,9 @@ char *taosCharsetReplace(char *charsetstr) { } /** + * TODO: here we may employ the systemctl API to set/get the correct locale on the Linux. In some cases, the setlocale + * seems does not response as expected. + * * In some Linux systems, setLocale(LC_CTYPE, "") may return NULL, in which case the launch of * both the TDengine Server and the Client may be interrupted. * @@ -148,7 +151,7 @@ void taosGetSystemLocale(char *outLocale, char *outCharset) { * * example: en_US.UTF-8, zh_CN.GB18030, zh_CN.UTF-8, * - * if user does not specify the locale in taos.cfg the program use default LC_CTYPE as system locale. + * If user does not specify the locale in taos.cfg, the program then uses default LC_CTYPE as system locale. * * In case of some CentOS systems, their default locale is "en_US.utf8", which is not valid code_page * for libiconv that is employed to convert string in this system. This program will automatically use diff --git a/source/os/src/osSysinfo.c b/source/os/src/osSysinfo.c index e5ca9faacb..51fff3a04f 100644 --- a/source/os/src/osSysinfo.c +++ b/source/os/src/osSysinfo.c @@ -155,8 +155,8 @@ static int32_t taosGetSysCpuInfo(SysCpuInfo *cpuInfo) { } char line[1024]; - ssize_t _bytes = taosGetsFile(pFile, sizeof(line), line); - if ((_bytes < 0) || (line == NULL)) { + ssize_t bytes = taosGetsFile(pFile, sizeof(line), line); + if (bytes < 0) { taosCloseFile(&pFile); return -1; } @@ -193,9 +193,9 @@ static int32_t taosGetProcCpuInfo(ProcCpuInfo *cpuInfo) { return -1; } - char line[1024]; - ssize_t _bytes = taosGetsFile(pFile, sizeof(line), line); - if ((_bytes < 0) || (line == NULL)) { + char line[1024] = {0}; + ssize_t bytes = taosGetsFile(pFile, sizeof(line), line); + if (bytes < 0) { taosCloseFile(&pFile); return -1; } @@ -239,6 +239,7 @@ void taosGetSystemInfo() { taosGetCpuCores(&tsNumOfCores); taosGetTotalMemory(&tsTotalMemoryKB); taosGetCpuUsage(NULL, NULL); + taosGetCpuInstructions(&tsSSE42Enable, &tsAVXEnable, &tsAVX2Enable, &tsFMAEnable); #endif } @@ -366,7 +367,7 @@ int32_t taosGetCpuInfo(char *cpuModel, int32_t maxLen, float *numOfCores) { return code; #else - char line[1024]; + char line[1024] = {0}; size_t size = 0; int32_t done = 0; int32_t code = -1; @@ -468,6 +469,46 @@ void taosGetCpuUsage(double *cpu_system, double *cpu_engine) { } } +#define __cpuid_fix(level, a, b, c, d) \ + __asm__("xor %%ecx, %%ecx\n" \ + "cpuid\n" \ + : "=a"(a), "=b"(b), "=c"(c), "=d"(d) \ + : "0"(level)) + +// todo add for windows and mac +int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma) { +#ifdef WINDOWS +#elif defined(_TD_DARWIN_64) +#else + + // Since the compiler is not support avx/avx2 instructions, the global variables always need to be + // set to be false +#if __AVX__ || __AVX2__ + tsSIMDEnable = true; +#else + tsSIMDEnable = false; +#endif + + uint32_t eax = 0, ebx = 0, ecx = 0, edx = 0; + + int32_t ret = __get_cpuid(1, &eax, &ebx, &ecx, &edx); + if (ret == 0) { + return -1; // failed to get the cpuid info + } + + *sse42 = (char) ((ecx & bit_SSE4_2) == bit_SSE4_2); + *avx = (char) ((ecx & bit_AVX) == bit_AVX); + *fma = (char) ((ecx & bit_FMA) == bit_FMA); + + // work around a bug in GCC. + // Ref to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77756 + __cpuid_fix(7u, eax, ebx, ecx, edx); + *avx2 = (char) ((ebx & bit_AVX2) == bit_AVX2); + return 0; + +#endif +} + int32_t taosGetTotalMemory(int64_t *totalKB) { #ifdef WINDOWS MEMORYSTATUSEX memsStat; @@ -511,11 +552,11 @@ int32_t taosGetProcMemory(int64_t *usedKB) { return -1; } - ssize_t _bytes = 0; - char line[1024]; + ssize_t bytes = 0; + char line[1024] = {0}; while (!taosEOFFile(pFile)) { - _bytes = taosGetsFile(pFile, sizeof(line), line); - if ((_bytes < 0) || (line == NULL)) { + bytes = taosGetsFile(pFile, sizeof(line), line); + if (bytes < 0) { break; } if (strstr(line, "VmRSS:") != NULL) { @@ -523,7 +564,7 @@ int32_t taosGetProcMemory(int64_t *usedKB) { } } - if (line == NULL) { + if (strlen(line) < 0) { // printf("read file:%s failed", tsProcMemFile); taosCloseFile(&pFile); return -1; @@ -624,14 +665,14 @@ int32_t taosGetProcIO(int64_t *rchars, int64_t *wchars, int64_t *read_bytes, int TdFilePtr pFile = taosOpenFile(tsProcIOFile, TD_FILE_READ | TD_FILE_STREAM); if (pFile == NULL) return -1; - ssize_t _bytes = 0; - char line[1024]; + ssize_t bytes = 0; + char line[1024] = {0}; char tmp[24]; int readIndex = 0; while (!taosEOFFile(pFile)) { - _bytes = taosGetsFile(pFile, sizeof(line), line); - if (_bytes < 10 || line == NULL) { + bytes = taosGetsFile(pFile, sizeof(line), line); + if (bytes < 10) { break; } if (strstr(line, "rchar:") != NULL) { diff --git a/source/os/src/osTime.c b/source/os/src/osTime.c index 58a09565f9..2771c8064f 100644 --- a/source/os/src/osTime.c +++ b/source/os/src/osTime.c @@ -339,7 +339,7 @@ char *taosStrpTime(const char *buf, const char *fmt, struct tm *tm) { #endif } -FORCE_INLINE int32_t taosGetTimeOfDay(struct timeval *tv) { +int32_t taosGetTimeOfDay(struct timeval *tv) { #ifdef WINDOWS time_t t; t = taosGetTimestampSec(); @@ -455,6 +455,7 @@ static int isLeapYear(time_t year) { else return 1; } + struct tm *taosLocalTimeNolock(struct tm *result, const time_t *timep, int dst) { if (result == NULL) { return localtime(timep); @@ -542,7 +543,9 @@ struct tm *taosLocalTimeNolock(struct tm *result, const time_t *timep, int dst) #endif return result; } + int32_t taosGetTimestampSec() { return (int32_t)time(NULL); } + int32_t taosClockGetTime(int clock_id, struct timespec *pTS) { #ifdef WINDOWS LARGE_INTEGER t; diff --git a/source/util/src/tconfig.c b/source/util/src/tconfig.c index c1fee37610..9949d9e4f1 100644 --- a/source/util/src/tconfig.c +++ b/source/util/src/tconfig.c @@ -561,13 +561,13 @@ void cfgDumpCfg(SConfig *pCfg, bool tsc, bool dump) { if (dump && strcmp(pItem->name, "scriptDir") == 0) continue; if (dump && strcmp(pItem->name, "simDebugFlag") == 0) continue; tstrncpy(src, cfgStypeStr(pItem->stype), CFG_SRC_PRINT_LEN); - for (int32_t i = 0; i < CFG_SRC_PRINT_LEN; ++i) { - if (src[i] == 0) src[i] = ' '; + for (int32_t j = 0; j < CFG_SRC_PRINT_LEN; ++j) { + if (src[j] == 0) src[j] = ' '; } tstrncpy(name, pItem->name, CFG_NAME_PRINT_LEN); - for (int32_t i = 0; i < CFG_NAME_PRINT_LEN; ++i) { - if (name[i] == 0) name[i] = ' '; + for (int32_t j = 0; j < CFG_NAME_PRINT_LEN; ++j) { + if (name[j] == 0) name[j] = ' '; } switch (pItem->dtype) { diff --git a/source/util/src/tcrc32c.c b/source/util/src/tcrc32c.c index bd662fa02c..795fe9dc4f 100644 --- a/source/util/src/tcrc32c.c +++ b/source/util/src/tcrc32c.c @@ -24,7 +24,6 @@ #endif #include "tcrc32c.h" -#include "tdef.h" #define POLY 0x82f63b78 #define LONG_SHIFT 8192 From f3be4454125a1d9683bb668ad4610a2fe9613eb1 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 11 Nov 2022 18:09:38 +0800 Subject: [PATCH 04/83] refactor: do some internal refactor. --- source/libs/executor/src/cachescanoperator.c | 25 +++++++++++++------- 1 file changed, 16 insertions(+), 9 deletions(-) diff --git a/source/libs/executor/src/cachescanoperator.c b/source/libs/executor/src/cachescanoperator.c index 6f9084ce52..5707c7cde0 100644 --- a/source/libs/executor/src/cachescanoperator.c +++ b/source/libs/executor/src/cachescanoperator.c @@ -30,6 +30,8 @@ static void destroyLastrowScanOperator(void* param); static int32_t extractCacheScanSlotId(const SArray* pColMatchInfo, SExecTaskInfo* pTaskInfo, int32_t** pSlotIds); static int32_t removeRedundantTsCol(SLastRowScanPhysiNode* pScanNode, SColMatchInfo* pColMatchInfo); +#define SCAN_ROW_TYPE(_t) ((_t)? CACHESCAN_RETRIEVE_LAST : CACHESCAN_RETRIEVE_LAST_ROW) + SOperatorInfo* createCacherowsScanOperator(SLastRowScanPhysiNode* pScanNode, SReadHandle* readHandle, SExecTaskInfo* pTaskInfo) { int32_t code = TSDB_CODE_SUCCESS; @@ -61,14 +63,14 @@ SOperatorInfo* createCacherowsScanOperator(SLastRowScanPhysiNode* pScanNode, SRe STableListInfo* pTableList = pTaskInfo->pTableInfoList; - initResultSizeInfo(&pOperator->resultInfo, 4096); - blockDataEnsureCapacity(pInfo->pRes, pOperator->resultInfo.capacity); + int32_t totalTables = tableListGetSize(pTableList); + int32_t capacity = 0; + pInfo->pUidList = taosArrayInit(4, sizeof(int64_t)); - // partition by tbname, todo opt perf - if (oneTableForEachGroup(pTableList) || (tableListGetSize(pTableList) == 1)) { - pInfo->retrieveType = - CACHESCAN_RETRIEVE_TYPE_ALL | (pScanNode->ignoreNull ? CACHESCAN_RETRIEVE_LAST : CACHESCAN_RETRIEVE_LAST_ROW); + // partition by tbname + if (oneTableForEachGroup(pTableList) || (totalTables == 1)) { + pInfo->retrieveType = CACHESCAN_RETRIEVE_TYPE_ALL | SCAN_ROW_TYPE(pScanNode->ignoreNull); STableKeyInfo* pList = tableListGetInfo(pTableList, 0); @@ -80,13 +82,18 @@ SOperatorInfo* createCacherowsScanOperator(SLastRowScanPhysiNode* pScanNode, SRe goto _error; } + capacity = TMIN(totalTables, 4096); + pInfo->pBufferredRes = createOneDataBlock(pInfo->pRes, false); - blockDataEnsureCapacity(pInfo->pBufferredRes, pOperator->resultInfo.capacity); + blockDataEnsureCapacity(pInfo->pBufferredRes, capacity); } else { // by tags - pInfo->retrieveType = CACHESCAN_RETRIEVE_TYPE_SINGLE | - (pScanNode->ignoreNull ? CACHESCAN_RETRIEVE_LAST : CACHESCAN_RETRIEVE_LAST_ROW); + pInfo->retrieveType = CACHESCAN_RETRIEVE_TYPE_SINGLE | SCAN_ROW_TYPE(pScanNode->ignoreNull); + capacity = 1; // only one row output } + initResultSizeInfo(&pOperator->resultInfo, capacity); + blockDataEnsureCapacity(pInfo->pRes, pOperator->resultInfo.capacity); + if (pScanNode->scan.pScanPseudoCols != NULL) { SExprSupp* p = &pInfo->pseudoExprSup; p->pExprInfo = createExprInfo(pScanNode->scan.pScanPseudoCols, NULL, &p->numOfExprs); From 6ae82b071e85c1d0902f72d65eff7e8c68f8f8e6 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 11 Nov 2022 18:28:16 +0800 Subject: [PATCH 05/83] refactor: do some internal refactor. --- source/libs/executor/src/sortoperator.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) diff --git a/source/libs/executor/src/sortoperator.c b/source/libs/executor/src/sortoperator.c index fc53623d44..9a5f6a031d 100644 --- a/source/libs/executor/src/sortoperator.c +++ b/source/libs/executor/src/sortoperator.c @@ -536,6 +536,7 @@ typedef struct SMultiwayMergeOperatorInfo { SSortHandle* pSortHandle; SColMatchInfo matchInfo; SSDataBlock* pInputBlock; + SSDataBlock* pIntermediateBlock; // to hold the intermediate result int64_t startTs; // sort start time bool groupSort; bool hasGroupId; @@ -635,12 +636,19 @@ SSDataBlock* getMultiwaySortedBlockData(SSortHandle* pHandle, SSDataBlock* pData SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo; blockDataCleanup(pDataBlock); - SSDataBlock* p = tsortGetSortedDataBlock(pHandle); - if (p == NULL) { - return NULL; + if (pInfo->pIntermediateBlock == NULL) { + pInfo->pIntermediateBlock = tsortGetSortedDataBlock(pHandle); + if (pInfo->pIntermediateBlock == NULL) { + terrno = TSDB_CODE_OUT_OF_MEMORY; + return NULL; + } + blockDataEnsureCapacity(pInfo->pIntermediateBlock, capacity); + } else { + blockDataCleanup(pInfo->pIntermediateBlock); } - blockDataEnsureCapacity(p, capacity); + SSDataBlock* p = pInfo->pIntermediateBlock; + while (1) { doGetSortedBlockData(pInfo, pHandle, capacity, p); if (p->info.rows == 0) { @@ -670,7 +678,6 @@ SSDataBlock* getMultiwaySortedBlockData(SSortHandle* pHandle, SSDataBlock* pData pDataBlock->info.groupId = pInfo->groupId; } - blockDataDestroy(p); qDebug("%s get sorted block, groupId:0x%" PRIx64 " rows:%d", GET_TASKID(pTaskInfo), pDataBlock->info.groupId, pDataBlock->info.rows); @@ -704,6 +711,7 @@ void destroyMultiwayMergeOperatorInfo(void* param) { SMultiwayMergeOperatorInfo* pInfo = (SMultiwayMergeOperatorInfo*)param; pInfo->binfo.pRes = blockDataDestroy(pInfo->binfo.pRes); pInfo->pInputBlock = blockDataDestroy(pInfo->pInputBlock); + pInfo->pIntermediateBlock = blockDataDestroy(pInfo->pIntermediateBlock); tsortDestroySortHandle(pInfo->pSortHandle); taosArrayDestroy(pInfo->pSortInfo); From ab52d28cb04558b29b3a72cf9c2e9c2ab9141396 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 11 Nov 2022 22:18:06 +0800 Subject: [PATCH 06/83] refactor: do some internal refactor. --- source/libs/executor/inc/executorimpl.h | 7 ++-- source/libs/executor/src/cachescanoperator.c | 3 +- source/libs/executor/src/exchangeoperator.c | 43 ++++++++++++-------- 3 files changed, 32 insertions(+), 21 deletions(-) diff --git a/source/libs/executor/inc/executorimpl.h b/source/libs/executor/inc/executorimpl.h index 62146b6048..e70cf37c63 100644 --- a/source/libs/executor/inc/executorimpl.h +++ b/source/libs/executor/inc/executorimpl.h @@ -254,17 +254,18 @@ typedef struct SExchangeInfo { SArray* pSourceDataInfo; tsem_t ready; void* pTransporter; + // SArray, result block list, used to keep the multi-block that // passed by downstream operator - SArray* pResultBlockList; - int32_t rspBlockIndex; // indicate the return block index in pResultBlockList + SArray* pReadyBlocks; + SArray* pRecycledBlocks;// build a pool for small data block to avoid to repeatly create and then destroy. SSDataBlock* pDummyBlock; // dummy block, not keep data bool seqLoadData; // sequential load data or not, false by default int32_t current; SLoadRemoteDataInfo loadInfo; uint64_t self; SLimitInfo limitInfo; - int64_t openedTs; // start exec time stamp + int64_t openedTs; // start exec time stamp, todo: move to SLoadRemoteDataInfo } SExchangeInfo; typedef struct SScanInfo { diff --git a/source/libs/executor/src/cachescanoperator.c b/source/libs/executor/src/cachescanoperator.c index 5707c7cde0..bbd46502ce 100644 --- a/source/libs/executor/src/cachescanoperator.c +++ b/source/libs/executor/src/cachescanoperator.c @@ -74,9 +74,8 @@ SOperatorInfo* createCacherowsScanOperator(SLastRowScanPhysiNode* pScanNode, SRe STableKeyInfo* pList = tableListGetInfo(pTableList, 0); - size_t num = tableListGetSize(pTableList); uint64_t suid = tableListGetSuid(pTableList); - code = tsdbCacherowsReaderOpen(pInfo->readHandle.vnode, pInfo->retrieveType, pList, num, + code = tsdbCacherowsReaderOpen(pInfo->readHandle.vnode, pInfo->retrieveType, pList, totalTables, taosArrayGetSize(pInfo->matchInfo.pList), suid, &pInfo->pLastrowReader); if (code != TSDB_CODE_SUCCESS) { goto _error; diff --git a/source/libs/executor/src/exchangeoperator.c b/source/libs/executor/src/exchangeoperator.c index 049de727df..4151018636 100644 --- a/source/libs/executor/src/exchangeoperator.c +++ b/source/libs/executor/src/exchangeoperator.c @@ -100,14 +100,23 @@ static void concurrentlyLoadRemoteDataImpl(SOperatorInfo* pOperator, SExchangeIn int32_t index = 0; char* pStart = pRetrieveRsp->data; while (index++ < pRetrieveRsp->numOfBlocks) { - SSDataBlock* pb = createOneDataBlock(pExchangeInfo->pDummyBlock, false); + SSDataBlock* pb = NULL; + + void* p = taosArrayPop(pExchangeInfo->pRecycledBlocks); + if (p != NULL) { + pb = *(SSDataBlock**) p; + blockDataCleanup(pb); + } else { + pb = createOneDataBlock(pExchangeInfo->pDummyBlock, false); + } + code = extractDataBlockFromFetchRsp(pb, pStart, NULL, &pStart); if (code != 0) { taosMemoryFreeClear(pDataInfo->pRsp); goto _error; } - taosArrayPush(pExchangeInfo->pResultBlockList, &pb); + taosArrayPush(pExchangeInfo->pReadyBlocks, &pb); } updateLoadRemoteInfo(pLoadInfo, pRetrieveRsp->numOfRows, pRetrieveRsp->compLen, pExchangeInfo->openedTs, pOperator); @@ -170,23 +179,26 @@ static SSDataBlock* doLoadRemoteDataImpl(SOperatorInfo* pOperator) { return NULL; } - size_t size = taosArrayGetSize(pExchangeInfo->pResultBlockList); - if (size == 0 || pExchangeInfo->rspBlockIndex >= size) { - pExchangeInfo->rspBlockIndex = 0; - taosArrayClearEx(pExchangeInfo->pResultBlockList, freeBlock); + // we have buffered retrieved datablock, return it directly + SSDataBlock** p = taosArrayPop(pExchangeInfo->pReadyBlocks); + if (p != NULL) { + taosArrayPush(pExchangeInfo->pRecycledBlocks, p); + return *p; + } else { if (pExchangeInfo->seqLoadData) { seqLoadRemoteData(pOperator); } else { concurrentlyLoadRemoteDataImpl(pOperator, pExchangeInfo, pTaskInfo); } - if (taosArrayGetSize(pExchangeInfo->pResultBlockList) == 0) { + if (taosArrayGetSize(pExchangeInfo->pReadyBlocks) == 0) { return NULL; + } else { + p = taosArrayPop(pExchangeInfo->pReadyBlocks); + taosArrayPush(pExchangeInfo->pRecycledBlocks, p); + return *p; } } - - // we have buffered retrieved datablock, return it directly - return taosArrayGetP(pExchangeInfo->pResultBlockList, pExchangeInfo->rspBlockIndex++); } static SSDataBlock* doLoadRemoteData(SOperatorInfo* pOperator) { @@ -284,7 +296,8 @@ SOperatorInfo* createExchangeOperatorInfo(void* pTransporter, SExchangePhysiNode tsem_init(&pInfo->ready, 0, 0); pInfo->pDummyBlock = createResDataBlock(pExNode->node.pOutputDataBlockDesc); - pInfo->pResultBlockList = taosArrayInit(1, POINTER_BYTES); + pInfo->pReadyBlocks = taosArrayInit(64, POINTER_BYTES); + pInfo->pRecycledBlocks = taosArrayInit(64, POINTER_BYTES); pInfo->seqLoadData = false; pInfo->pTransporter = pTransporter; @@ -326,11 +339,9 @@ void doDestroyExchangeOperatorInfo(void* param) { taosArrayDestroy(pExInfo->pSources); taosArrayDestroyEx(pExInfo->pSourceDataInfo, freeSourceDataInfo); - if (pExInfo->pResultBlockList != NULL) { - taosArrayDestroyEx(pExInfo->pResultBlockList, freeBlock); - pExInfo->pResultBlockList = NULL; - } - + taosArrayDestroyEx(pExInfo->pReadyBlocks, freeBlock); + taosArrayDestroyEx(pExInfo->pRecycledBlocks, freeBlock); + blockDataDestroy(pExInfo->pDummyBlock); tsem_destroy(&pExInfo->ready); From 7b53b8142e40c1750eb82ad2b76116e467ae50e4 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 12 Nov 2022 16:03:47 +0800 Subject: [PATCH 07/83] refactor: do some internal refactor. --- source/common/src/tdatablock.c | 1 + source/libs/executor/inc/executorimpl.h | 18 +++++++++--------- source/libs/executor/src/executor.c | 16 ++++++++++++++-- source/libs/executor/src/executorimpl.c | 7 +++++++ source/libs/qworker/src/qworker.c | 7 +------ 5 files changed, 32 insertions(+), 17 deletions(-) diff --git a/source/common/src/tdatablock.c b/source/common/src/tdatablock.c index 536cbed33e..f5dba35440 100644 --- a/source/common/src/tdatablock.c +++ b/source/common/src/tdatablock.c @@ -1306,6 +1306,7 @@ int32_t copyDataBlock(SSDataBlock* dst, const SSDataBlock* src) { colDataAssign(pDst, pSrc, src->info.rows, &src->info); } + uint32_t cap = dst->info.capacity; dst->info = src->info; dst->info.capacity = cap; diff --git a/source/libs/executor/inc/executorimpl.h b/source/libs/executor/inc/executorimpl.h index e70cf37c63..a2e269dd51 100644 --- a/source/libs/executor/inc/executorimpl.h +++ b/source/libs/executor/inc/executorimpl.h @@ -153,14 +153,13 @@ typedef struct { SSchemaWrapper* qsw; } SSchemaInfo; -typedef struct SExecTaskInfo { - STaskIdInfo id; - uint32_t status; - STimeWindow window; - STaskCostInfo cost; - int64_t owner; // if it is in execution - int32_t code; - +struct SExecTaskInfo { + STaskIdInfo id; + uint32_t status; + STimeWindow window; + STaskCostInfo cost; + int64_t owner; // if it is in execution + int32_t code; int64_t version; // used for stream to record wal version SStreamTaskInfo streamInfo; SSchemaInfo schemaInfo; @@ -171,7 +170,8 @@ typedef struct SExecTaskInfo { SSubplan* pSubplan; struct SOperatorInfo* pRoot; SLocalFetch localFetch; -} SExecTaskInfo; + SArray* pResultBlockList;// result block list +}; enum { OP_NOT_OPENED = 0x0, diff --git a/source/libs/executor/src/executor.c b/source/libs/executor/src/executor.c index 1aa9a3c613..0163a389ef 100644 --- a/source/libs/executor/src/executor.c +++ b/source/libs/executor/src/executor.c @@ -536,7 +536,7 @@ int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, bo memcpy(&pTaskInfo->localFetch, pLocal, sizeof(*pLocal)); } - taosArrayClearEx(pResList, freeBlock); + taosArrayClear(pResList); int64_t curOwner = 0; if ((curOwner = atomic_val_compare_exchange_64(&pTaskInfo->owner, 0, threadId)) != 0) { @@ -574,8 +574,20 @@ int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, bo int64_t st = taosGetTimestampUs(); + int32_t blockIndex = 0; while ((pRes = pTaskInfo->pRoot->fpSet.getNextFn(pTaskInfo->pRoot)) != NULL) { - SSDataBlock* p = createOneDataBlock(pRes, true); + SSDataBlock* p = NULL; + if (blockIndex >= taosArrayGetSize(pTaskInfo->pResultBlockList)) { + SSDataBlock* p1 = createOneDataBlock(pRes, true); + taosArrayPush(pTaskInfo->pResultBlockList, &p1); + p = p1; + } else { + p = *(SSDataBlock**) taosArrayGet(pTaskInfo->pResultBlockList, blockIndex); + copyDataBlock(p, pRes); + } + + blockIndex += 1; + current += p->info.rows; ASSERT(p->info.rows > 0); taosArrayPush(pResList, &p); diff --git a/source/libs/executor/src/executorimpl.c b/source/libs/executor/src/executorimpl.c index 709e981a1f..f164622687 100644 --- a/source/libs/executor/src/executorimpl.c +++ b/source/libs/executor/src/executorimpl.c @@ -2600,6 +2600,7 @@ static SExecTaskInfo* createExecTaskInfo(uint64_t queryId, uint64_t taskId, EOPT pTaskInfo->id.queryId = queryId; pTaskInfo->execModel = model; pTaskInfo->pTableInfoList = tableListCreate(); + pTaskInfo->pResultBlockList = taosArrayInit(128, POINTER_BYTES); char* p = taosMemoryCalloc(1, 128); snprintf(p, 128, "TID:0x%" PRIx64 " QID:0x%" PRIx64, taskId, queryId); @@ -3201,6 +3202,11 @@ _complete: return terrno; } +static void freeBlock(void* pParam) { + SSDataBlock* pBlock = *(SSDataBlock**)pParam; + blockDataDestroy(pBlock); +} + void doDestroyTask(SExecTaskInfo* pTaskInfo) { qDebug("%s execTask is freed", GET_TASKID(pTaskInfo)); @@ -3213,6 +3219,7 @@ void doDestroyTask(SExecTaskInfo* pTaskInfo) { nodesDestroyNode((SNode*)pTaskInfo->pSubplan); } + taosArrayDestroyEx(pTaskInfo->pResultBlockList, freeBlock); taosMemoryFreeClear(pTaskInfo->sql); taosMemoryFreeClear(pTaskInfo->id.str); taosMemoryFreeClear(pTaskInfo); diff --git a/source/libs/qworker/src/qworker.c b/source/libs/qworker/src/qworker.c index a7cd3db824..2606556838 100644 --- a/source/libs/qworker/src/qworker.c +++ b/source/libs/qworker/src/qworker.c @@ -18,11 +18,6 @@ SQWorkerMgmt gQwMgmt = { .qwNum = 0, }; -static void freeBlock(void *param) { - SSDataBlock *pBlock = *(SSDataBlock **)param; - blockDataDestroy(pBlock); -} - int32_t qwProcessHbLinkBroken(SQWorker *mgmt, SQWMsg *qwMsg, SSchedulerHbReq *req) { int32_t code = 0; SSchedulerHbRsp rsp = {0}; @@ -193,7 +188,7 @@ int32_t qwExecTask(QW_FPARAMS_DEF, SQWTaskCtx *ctx, bool *queryStop) { } _return: - taosArrayDestroyEx(pResList, freeBlock); + taosArrayDestroy(pResList); QW_RET(code); } From c888cbf068dd09103256a4e8061fa5d57290078e Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Mon, 14 Nov 2022 14:14:24 +0800 Subject: [PATCH 08/83] refactor: do some internal refactor. --- include/libs/function/function.h | 2 +- source/libs/executor/src/executorimpl.c | 16 +- source/libs/function/src/builtinsimpl.c | 30 +- .../libs/function/src/detail/tavgfunction.c | 300 ++++++++++++++++-- source/libs/function/src/detail/tminmax.c | 6 +- 5 files changed, 304 insertions(+), 50 deletions(-) diff --git a/include/libs/function/function.h b/include/libs/function/function.h index 6f2a675466..240772bfc2 100644 --- a/include/libs/function/function.h +++ b/include/libs/function/function.h @@ -115,7 +115,7 @@ typedef struct SInputColumnInfoData { int32_t startRowIndex; // handle started row index int32_t numOfRows; // the number of rows needs to be handled int32_t numOfInputCols; // PTS is not included - bool colDataAggIsSet; // if agg is set or not + bool colDataSMAIsSet; // if agg is set or not SColumnInfoData *pPTS; // primary timestamp column SColumnInfoData **pData; SColumnDataAgg **pColumnDataAgg; diff --git a/source/libs/executor/src/executorimpl.c b/source/libs/executor/src/executorimpl.c index f164622687..5031d75231 100644 --- a/source/libs/executor/src/executorimpl.c +++ b/source/libs/executor/src/executorimpl.c @@ -349,13 +349,13 @@ typedef struct { } SFunctionCtxStatus; static void functionCtxSave(SqlFunctionCtx* pCtx, SFunctionCtxStatus* pStatus) { - pStatus->hasAgg = pCtx->input.colDataAggIsSet; + pStatus->hasAgg = pCtx->input.colDataSMAIsSet; pStatus->numOfRows = pCtx->input.numOfRows; pStatus->startOffset = pCtx->input.startRowIndex; } static void functionCtxRestore(SqlFunctionCtx* pCtx, SFunctionCtxStatus* pStatus) { - pCtx->input.colDataAggIsSet = pStatus->hasAgg; + pCtx->input.colDataSMAIsSet = pStatus->hasAgg; pCtx->input.numOfRows = pStatus->numOfRows; pCtx->input.startRowIndex = pStatus->startOffset; } @@ -372,8 +372,8 @@ void doApplyFunctions(SExecTaskInfo* taskInfo, SqlFunctionCtx* pCtx, SColumnInfo // not a whole block involved in query processing, statistics data can not be used // NOTE: the original value of isSet have been changed here - if (pCtx[k].input.colDataAggIsSet && forwardStep < numOfTotal) { - pCtx[k].input.colDataAggIsSet = false; + if (pCtx[k].input.colDataSMAIsSet && forwardStep < numOfTotal) { + pCtx[k].input.colDataSMAIsSet = false; } if (fmIsWindowPseudoColumnFunc(pCtx[k].functionId)) { @@ -486,7 +486,7 @@ static int32_t doSetInputDataBlock(SExprSupp* pExprSup, SSDataBlock* pBlock, int SInputColumnInfoData* pInput = &pCtx[i].input; pInput->uid = pBlock->info.uid; - pInput->colDataAggIsSet = false; + pInput->colDataSMAIsSet = false; SExprInfo* pOneExpr = &pExprSup->pExprInfo[i]; for (int32_t j = 0; j < pOneExpr->base.numOfParams; ++j) { @@ -798,7 +798,7 @@ void setBlockSMAInfo(SqlFunctionCtx* pCtx, SExprInfo* pExprInfo, SSDataBlock* pB pInput->totalRows = numOfRows; if (pBlock->pBlockAgg != NULL) { - pInput->colDataAggIsSet = true; + pInput->colDataSMAIsSet = true; for (int32_t j = 0; j < pExprInfo->base.numOfParams; ++j) { SFunctParam* pFuncParam = &pExprInfo->base.pParam[j]; @@ -807,7 +807,7 @@ void setBlockSMAInfo(SqlFunctionCtx* pCtx, SExprInfo* pExprInfo, SSDataBlock* pB int32_t slotId = pFuncParam->pCol->slotId; pInput->pColumnDataAgg[j] = pBlock->pBlockAgg[slotId]; if (pInput->pColumnDataAgg[j] == NULL) { - pInput->colDataAggIsSet = false; + pInput->colDataSMAIsSet = false; } // Here we set the column info data since the data type for each column data is required, but @@ -818,7 +818,7 @@ void setBlockSMAInfo(SqlFunctionCtx* pCtx, SExprInfo* pExprInfo, SSDataBlock* pB } } } else { - pInput->colDataAggIsSet = false; + pInput->colDataSMAIsSet = false; } } diff --git a/source/libs/function/src/builtinsimpl.c b/source/libs/function/src/builtinsimpl.c index 26f9c3ad0b..bf79cb5191 100644 --- a/source/libs/function/src/builtinsimpl.c +++ b/source/libs/function/src/builtinsimpl.c @@ -498,13 +498,13 @@ static int32_t getNumOfElems(SqlFunctionCtx* pCtx) { int32_t numOfElem = 0; /* - * 1. column data missing (schema modified) causes pInputCol->hasNull == true. pInput->colDataAggIsSet == true; - * 2. for general non-primary key columns, pInputCol->hasNull may be true or false, pInput->colDataAggIsSet == true; - * 3. for primary key column, pInputCol->hasNull always be false, pInput->colDataAggIsSet == false; + * 1. column data missing (schema modified) causes pInputCol->hasNull == true. pInput->colDataSMAIsSet == true; + * 2. for general non-primary key columns, pInputCol->hasNull may be true or false, pInput->colDataSMAIsSet == true; + * 3. for primary key column, pInputCol->hasNull always be false, pInput->colDataSMAIsSet == false; */ SInputColumnInfoData* pInput = &pCtx->input; SColumnInfoData* pInputCol = pInput->pData[0]; - if (pInput->colDataAggIsSet && pInput->totalRows == pInput->numOfRows) { + if (pInput->colDataSMAIsSet && pInput->totalRows == pInput->numOfRows) { numOfElem = pInput->numOfRows - pInput->pColumnDataAgg[0]->numOfNull; ASSERT(numOfElem >= 0); } else { @@ -593,7 +593,7 @@ int32_t sumFunction(SqlFunctionCtx* pCtx) { goto _sum_over; } - if (pInput->colDataAggIsSet) { + if (pInput->colDataSMAIsSet) { numOfElem = pInput->numOfRows - pAgg->numOfNull; ASSERT(numOfElem >= 0); @@ -658,7 +658,7 @@ int32_t sumInvertFunction(SqlFunctionCtx* pCtx) { SSumRes* pSumRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - if (pInput->colDataAggIsSet) { + if (pInput->colDataSMAIsSet) { numOfElem = pInput->numOfRows - pAgg->numOfNull; ASSERT(numOfElem >= 0); @@ -770,7 +770,7 @@ bool getSumFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { // goto _avg_over; // } // -// if (pInput->colDataAggIsSet) { +// if (pInput->colDataSMAIsSet) { // numOfElem = numOfRows - pAgg->numOfNull; // ASSERT(numOfElem >= 0); // @@ -1161,7 +1161,7 @@ bool getMinmaxFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { // } // // // data in current data block are qualified to the query -// if (pInput->colDataAggIsSet) { +// if (pInput->colDataSMAIsSet) { // numOfElems = pInput->numOfRows - pAgg->numOfNull; // ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); // if (numOfElems == 0) { @@ -2471,7 +2471,7 @@ int32_t percentileFunction(SqlFunctionCtx* pCtx) { // the first stage, only acquire the min/max value if (pInfo->stage == 0) { - if (pCtx->input.colDataAggIsSet) { + if (pCtx->input.colDataSMAIsSet) { double tmin = 0.0, tmax = 0.0; if (IS_SIGNED_NUMERIC_TYPE(type)) { tmin = (double)GET_INT64_VAL(&pAgg->min); @@ -2933,14 +2933,14 @@ int32_t firstFunction(SqlFunctionCtx* pCtx) { pInfo->bytes = pInputCol->info.bytes; // All null data column, return directly. - if (pInput->colDataAggIsSet && (pInput->pColumnDataAgg[0]->numOfNull == pInput->totalRows)) { + if (pInput->colDataSMAIsSet && (pInput->pColumnDataAgg[0]->numOfNull == pInput->totalRows)) { ASSERT(pInputCol->hasNull == true); // save selectivity value for column consisted of all null values firstlastSaveTupleData(pCtx->pSrcBlock, pInput->startRowIndex, pCtx, pInfo); return 0; } - SColumnDataAgg* pColAgg = (pInput->colDataAggIsSet) ? pInput->pColumnDataAgg[0] : NULL; + SColumnDataAgg* pColAgg = (pInput->colDataSMAIsSet) ? pInput->pColumnDataAgg[0] : NULL; TSKEY startKey = getRowPTs(pInput->pPTS, 0); TSKEY endKey = getRowPTs(pInput->pPTS, pInput->totalRows - 1); @@ -3037,14 +3037,14 @@ int32_t lastFunction(SqlFunctionCtx* pCtx) { pInfo->bytes = bytes; // All null data column, return directly. - if (pInput->colDataAggIsSet && (pInput->pColumnDataAgg[0]->numOfNull == pInput->totalRows)) { + if (pInput->colDataSMAIsSet && (pInput->pColumnDataAgg[0]->numOfNull == pInput->totalRows)) { ASSERT(pInputCol->hasNull == true); // save selectivity value for column consisted of all null values firstlastSaveTupleData(pCtx->pSrcBlock, pInput->startRowIndex, pCtx, pInfo); return 0; } - SColumnDataAgg* pColAgg = (pInput->colDataAggIsSet) ? pInput->pColumnDataAgg[0] : NULL; + SColumnDataAgg* pColAgg = (pInput->colDataSMAIsSet) ? pInput->pColumnDataAgg[0] : NULL; TSKEY startKey = getRowPTs(pInput->pPTS, 0); TSKEY endKey = getRowPTs(pInput->pPTS, pInput->totalRows - 1); @@ -3988,7 +3988,7 @@ int32_t spreadFunction(SqlFunctionCtx* pCtx) { SSpreadInfo* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - if (pInput->colDataAggIsSet) { + if (pInput->colDataSMAIsSet) { numOfElems = pInput->numOfRows - pAgg->numOfNull; if (numOfElems == 0) { goto _spread_over; @@ -4163,7 +4163,7 @@ int32_t elapsedFunction(SqlFunctionCtx* pCtx) { goto _elapsed_over; } - if (pInput->colDataAggIsSet) { + if (pInput->colDataSMAIsSet) { if (pInfo->min == TSKEY_MAX) { pInfo->min = GET_INT64_VAL(&pAgg->min); pInfo->max = GET_INT64_VAL(&pAgg->max); diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 01e0a499eb..9c3b9cf573 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -48,15 +48,14 @@ typedef struct SAvgRes { int16_t type; // store the original input type, used in merge function } SAvgRes; -static void floatVectorSumAVX(const SInputColumnInfoData* pInput, const float* plist, SAvgRes* pRes) { +static void floatVectorSumAVX(const float* plist, int32_t numOfRows, SAvgRes* pRes) { #if __AVX__ // find the start position that are aligned to 32bytes address in memory - int32_t startIndex = 0; //((uint64_t)plist) & ((1<<8u)-1); int32_t bitWidth = 8; + int32_t remainder = numOfRows % bitWidth; + int32_t rounds = numOfRows / bitWidth; - int32_t remain = (pInput->numOfRows - startIndex) % bitWidth; - int32_t rounds = (pInput->numOfRows - startIndex) / bitWidth; - const float* p = &plist[startIndex]; + const float* p = plist; __m256 val; __m256 sum = _mm256_setzero_ps(); @@ -71,18 +70,126 @@ static void floatVectorSumAVX(const SInputColumnInfoData* pInput, const float* p const float* q = (const float*)∑ pRes->sum.dsum += q[0] + q[1] + q[2] + q[3] + q[4] + q[5] + q[6] + q[7]; - // calculate the front and the reminder items in array list - for (int32_t j = 0; j < startIndex; ++j) { - pRes->sum.dsum += plist[j]; + int32_t startIndex = rounds * bitWidth; + for (int32_t j = 0; j < remainder; ++j) { + pRes->sum.dsum += plist[j + startIndex]; + } +#endif +} + +static void doubleVectorSumAVX(const double* plist, int32_t numOfRows, SAvgRes* pRes) { +#if __AVX__ + // find the start position that are aligned to 32bytes address in memory + int32_t bitWidth = 4; + int32_t remainder = numOfRows % bitWidth; + int32_t rounds = numOfRows / bitWidth; + + const double* p = plist; + + __m256d val; + __m256d sum = _mm256_setzero_pd(); + + for (int32_t i = 0; i < rounds; ++i) { + val = _mm256_loadu_pd(p); + sum = _mm256_add_pd(sum, val); + p += bitWidth; } - startIndex += rounds * bitWidth; - for (int32_t j = 0; j < remain; ++j) { + // let sum up the final results + const double* q = (const double*)∑ + pRes->sum.dsum += q[0] + q[1] + q[2] + q[3]; + + int32_t startIndex = rounds * bitWidth; + for (int32_t j = 0; j < remainder; ++j) { pRes->sum.dsum += plist[j + startIndex]; } #endif } +static void i8VectorSumAVX2(const int8_t* plist, int32_t numOfRows, SAvgRes* pRes) { +#if __AVX2__ + // find the start position that are aligned to 32bytes address in memory + int32_t bitWidth = 16; + int32_t remainder = numOfRows % bitWidth; + int32_t rounds = numOfRows / bitWidth; + + const int8_t* p = plist; + + __m256i sum = _mm256_setzero_si256(); + + for (int32_t i = 0; i < rounds; ++i) { + __m256i val = _mm256_lddqu_si256((__m256i*)p); +// __m256i extVal = _mm256_cvtepi8_epi64(val); + sum = _mm256_add_epi8(sum, val); + p += bitWidth; + } + + // let sum up the final results + const int8_t* q = (const int8_t*)∑ + pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; + + int32_t startIndex = rounds * bitWidth; + for (int32_t j = 0; j < remainder; ++j) { + pRes->sum.isum += plist[j + startIndex]; + } +#endif +} + +static void i32VectorSumAVX2(const int32_t* plist, int32_t numOfRows, SAvgRes* pRes) { +#if __AVX2__ + // find the start position that are aligned to 32bytes address in memory + int32_t bitWidth = 8; + int32_t remainder = numOfRows % bitWidth; + int32_t rounds = numOfRows / bitWidth; + + const int32_t* p = plist; + + __m256i sum = _mm256_setzero_si256(); + for (int32_t i = 0; i < rounds; ++i) { + __m256i val = _mm256_lddqu_si256((__m256i*)p); + sum = _mm256_add_epi32(sum, val); + p += bitWidth; + } + + // let sum up the final results + const int64_t* q = (const int64_t*)∑ + pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; + + int32_t startIndex = rounds * bitWidth; + for (int32_t j = 0; j < remainder; ++j) { + pRes->sum.isum += plist[j + startIndex]; + } +#endif +} + +static void i64VectorSumAVX2(const int64_t* plist, int32_t numOfRows, SAvgRes* pRes) { +#if __AVX2__ + // find the start position that are aligned to 32bytes address in memory + int32_t bitWidth = 4; + int32_t remainder = numOfRows % bitWidth; + int32_t rounds = numOfRows / bitWidth; + + const int64_t* p = plist; + + __m256i sum = _mm256_setzero_si256(); + + for (int32_t i = 0; i < rounds; ++i) { + __m256i val = _mm256_lddqu_si256((__m256i*)p); + sum = _mm256_add_epi64(sum, val); + p += bitWidth; + } + + // let sum up the final results + const int64_t* q = (const int64_t*)∑ + pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; + + int32_t startIndex = rounds * bitWidth; + for (int32_t j = 0; j < remainder; ++j) { + pRes->sum.isum += plist[j + startIndex]; + } +#endif +} + static int32_t handleFloatCols(const SColumnInfoData* pCol, const SInputColumnInfoData* pInput, SAvgRes* pRes) { int32_t numOfElems = 0; float* plist = (float*)pCol->pData; @@ -105,7 +212,7 @@ static int32_t handleFloatCols(const SColumnInfoData* pCol, const SInputColumnIn // 3. If the CPU supports AVX, let's employ AVX instructions to speedup this loop if (tsAVXEnable && tsSIMDEnable) { - floatVectorSumAVX(pInput, plist, pRes); + floatVectorSumAVX(plist, pInput->numOfRows, pRes); } else { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { pRes->sum.dsum += plist[i]; @@ -133,8 +240,25 @@ bool avgFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo) { return true; } +static int32_t calculateAvgBySMAInfo(SAvgRes* pRes, int32_t numOfRows, int32_t type, const SColumnDataAgg* pAgg) { + int32_t numOfElem = numOfRows - pAgg->numOfNull; + ASSERT(numOfElem >= 0); + + pRes->count += numOfElem; + if (IS_SIGNED_NUMERIC_TYPE(type)) { + pRes->sum.isum += pAgg->sum; + } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { + pRes->sum.usum += pAgg->sum; + } else if (IS_FLOAT_TYPE(type)) { + pRes->sum.dsum += GET_DOUBLE_VAL((const char*)&(pAgg->sum)); + } + + return numOfElem; +} + int32_t avgFunction(SqlFunctionCtx* pCtx) { - int32_t numOfElem = 0; + int32_t numOfElem = 0; + const int32_t THRESHOLD_SIZE = 8; SInputColumnInfoData* pInput = &pCtx->input; SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; @@ -154,19 +278,149 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { goto _avg_over; } - if (pInput->colDataAggIsSet) { - numOfElem = numOfRows - pAgg->numOfNull; - ASSERT(numOfElem >= 0); + if (pInput->colDataSMAIsSet) { // try to use SMA if available + numOfElem = calculateAvgBySMAInfo(pAvgRes, numOfRows, type, pAgg); + } else if (!pCol->hasNull) { // try to employ the simd instructions to speed up the loop + numOfElem = pInput->numOfRows; + pAvgRes->count += pInput->numOfRows; - pAvgRes->count += numOfElem; - if (IS_SIGNED_NUMERIC_TYPE(type)) { - pAvgRes->sum.isum += pAgg->sum; - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - pAvgRes->sum.usum += pAgg->sum; - } else if (IS_FLOAT_TYPE(type)) { - pAvgRes->sum.dsum += GET_DOUBLE_VAL((const char*)&(pAgg->sum)); + bool simdAvaiable = tsAVXEnable && tsSIMDEnable && (numOfRows > THRESHOLD_SIZE); + + switch(type) { + case TSDB_DATA_TYPE_TINYINT: { + const int8_t* plist = (const int8_t*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + i8VectorSumAVX2(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.isum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_SMALLINT: { + const double* plist = (const double*)pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + doubleVectorSumAVX(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.isum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_INT: { + const int32_t* plist = (const int32_t*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + i32VectorSumAVX2(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.isum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_BIGINT: { + const int64_t* plist = (const int64_t*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + i64VectorSumAVX2(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.isum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_FLOAT: { + const float* plist = (const float*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + floatVectorSumAVX(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.dsum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_DOUBLE: { + const double* plist = (const double*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + doubleVectorSumAVX(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.dsum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_UTINYINT: { + const double* plist = (const double*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + doubleVectorSumAVX(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.usum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_USMALLINT: { + const double* plist = (const double*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + doubleVectorSumAVX(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.usum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_UINT: { + const double* plist = (const double*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + doubleVectorSumAVX(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.usum += plist[i]; + } + } + break; + } + case TSDB_DATA_TYPE_UBIGINT: { + const double* plist = (const double*) pCol->pData; + + // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop + if (simdAvaiable) { + doubleVectorSumAVX(plist, numOfRows, pAvgRes); + } else { + for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { + pAvgRes->sum.usum += plist[i]; + } + } + break; + } + default: + ASSERT(0); } - } else { // computing based on the true data block + } else { switch (type) { case TSDB_DATA_TYPE_TINYINT: { int8_t* plist = (int8_t*)pCol->pData; diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index 074e5ef428..d239315e0e 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -36,7 +36,7 @@ static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool is if (!isMinFunc) { // max function for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_loadu_si256((__m256i*)p); + next = _mm256_lddqu_si256((__m256i*)p); initialVal = _mm256_max_epi32(initialVal, next); p += bitWidth; } @@ -61,7 +61,7 @@ static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool is } } else { // min function for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_loadu_si256((__m256i*)p); + next = _mm256_lddqu_si256((__m256i*)p); initialVal = _mm256_min_epi32(initialVal, next); p += bitWidth; } @@ -369,7 +369,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { } // data in current data block are qualified to the query - if (pInput->colDataAggIsSet) { + if (pInput->colDataSMAIsSet) { numOfElems = pInput->numOfRows - pAgg->numOfNull; ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); if (numOfElems == 0) { From 618b4ab8810ea883224d869035c75a6af8bf542b Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Mon, 14 Nov 2022 14:15:48 +0800 Subject: [PATCH 09/83] refactor: do some internal refactor. --- source/libs/function/src/detail/tavgfunction.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 9c3b9cf573..8c26886329 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -152,8 +152,8 @@ static void i32VectorSumAVX2(const int32_t* plist, int32_t numOfRows, SAvgRes* p } // let sum up the final results - const int64_t* q = (const int64_t*)∑ - pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; + const int32_t* q = (const int32_t*)∑ + pRes->sum.isum += q[0] + q[1] + q[2] + q[3] + q[4] + q[5] + q[6] + q[7]; int32_t startIndex = rounds * bitWidth; for (int32_t j = 0; j < remainder; ++j) { From c35f668c17efc567a1e6d25fb13bd7bc0cdfcb91 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Mon, 14 Nov 2022 14:37:28 +0800 Subject: [PATCH 10/83] refactor: increase the initial buffer size. --- source/libs/executor/src/executil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index f0db51dc9d..44390ca2e5 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1770,7 +1770,7 @@ STableListInfo* tableListCreate() { goto _error; } - pListInfo->map = taosHashInit(32, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_ENTRY_LOCK); + pListInfo->map = taosHashInit(1024, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_ENTRY_LOCK); if (pListInfo->map == NULL) { goto _error; } From 8b0b351d3915bdeff511d02bfd74e38f99e680e2 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Tue, 15 Nov 2022 09:59:20 +0800 Subject: [PATCH 11/83] refactor: do some internal refactor. --- cmake/cmake.define | 4 +- cmake/cmake.platform | 4 +- include/util/thash.h | 2 +- .../libs/function/src/detail/tavgfunction.c | 564 ++++++++++-------- source/util/src/thash.c | 6 +- 5 files changed, 311 insertions(+), 269 deletions(-) diff --git a/cmake/cmake.define b/cmake/cmake.define index 3b6024efc8..e34785cba6 100644 --- a/cmake/cmake.define +++ b/cmake/cmake.define @@ -134,9 +134,9 @@ ELSE () IF("${SIMD_SUPPORT}" MATCHES "true") ADD_DEFINITIONS("-mavx -mavx2") - MESSAGE(STATUS "cpu simd instruction AVX/AVX2 supported") + MESSAGE(STATUS "SIMD instructions (AVX/AVX2) is ACTIVATED") ELSE() - MESSAGE(STATUS "cpu simd instruction AVX/AVX2 NOT supported") + MESSAGE(STATUS "SIMD instruction (AVX/AVX2)is NOT ACTIVATED") ENDIF() ENDIF () diff --git a/cmake/cmake.platform b/cmake/cmake.platform index c3680e0de4..3623925a2b 100644 --- a/cmake/cmake.platform +++ b/cmake/cmake.platform @@ -135,5 +135,5 @@ ENDIF () MESSAGE(STATUS "platform arch:" ${PLATFORM_ARCH_STR}) -MESSAGE("C Compiler ID: ${CMAKE_C_COMPILER_ID}") -MESSAGE("CXX Compiler ID: ${CMAKE_CXX_COMPILER_ID}") +MESSAGE("C Compiler: ${CMAKE_C_COMPILER} (${CMAKE_C_COMPILER_ID}, ${CMAKE_C_COMPILER_VERSION})") +MESSAGE("CXX Compiler: ${CMAKE_CXX_COMPILER} (${CMAKE_C_COMPILER_ID}, ${CMAKE_CXX_COMPILER_VERSION})") diff --git a/include/util/thash.h b/include/util/thash.h index 08caad495d..a04f78a3d1 100644 --- a/include/util/thash.h +++ b/include/util/thash.h @@ -213,7 +213,7 @@ void taosHashSetEqualFp(SHashObj *pHashObj, _equal_fn_t fp); */ void taosHashSetFreeFp(SHashObj *pHashObj, _hash_free_fn_t fp); -int64_t taosHashGetCompTimes(SHashObj *pHashObj); +//int64_t taosHashGetCompTimes(SHashObj *pHashObj); #ifdef __cplusplus } diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 8c26886329..d7ef73b08e 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -49,11 +49,14 @@ typedef struct SAvgRes { } SAvgRes; static void floatVectorSumAVX(const float* plist, int32_t numOfRows, SAvgRes* pRes) { + const int32_t bitWidth = 256; + #if __AVX__ // find the start position that are aligned to 32bytes address in memory - int32_t bitWidth = 8; - int32_t remainder = numOfRows % bitWidth; - int32_t rounds = numOfRows / bitWidth; + int32_t width = (bitWidth>>3u) / sizeof(float); + + int32_t remainder = numOfRows % width; + int32_t rounds = numOfRows / width; const float* p = plist; @@ -63,14 +66,14 @@ static void floatVectorSumAVX(const float* plist, int32_t numOfRows, SAvgRes* pR for (int32_t i = 0; i < rounds; ++i) { val = _mm256_loadu_ps(p); sum = _mm256_add_ps(sum, val); - p += bitWidth; + p += width; } // let sum up the final results const float* q = (const float*)∑ pRes->sum.dsum += q[0] + q[1] + q[2] + q[3] + q[4] + q[5] + q[6] + q[7]; - int32_t startIndex = rounds * bitWidth; + int32_t startIndex = rounds * width; for (int32_t j = 0; j < remainder; ++j) { pRes->sum.dsum += plist[j + startIndex]; } @@ -78,11 +81,14 @@ static void floatVectorSumAVX(const float* plist, int32_t numOfRows, SAvgRes* pR } static void doubleVectorSumAVX(const double* plist, int32_t numOfRows, SAvgRes* pRes) { + const int32_t bitWidth = 256; + #if __AVX__ // find the start position that are aligned to 32bytes address in memory - int32_t bitWidth = 4; - int32_t remainder = numOfRows % bitWidth; - int32_t rounds = numOfRows / bitWidth; + int32_t width = (bitWidth>>3u) / sizeof(int64_t); + + int32_t remainder = numOfRows % width; + int32_t rounds = numOfRows / width; const double* p = plist; @@ -92,70 +98,143 @@ static void doubleVectorSumAVX(const double* plist, int32_t numOfRows, SAvgRes* for (int32_t i = 0; i < rounds; ++i) { val = _mm256_loadu_pd(p); sum = _mm256_add_pd(sum, val); - p += bitWidth; + p += width; } // let sum up the final results const double* q = (const double*)∑ pRes->sum.dsum += q[0] + q[1] + q[2] + q[3]; - int32_t startIndex = rounds * bitWidth; + int32_t startIndex = rounds * width; for (int32_t j = 0; j < remainder; ++j) { pRes->sum.dsum += plist[j + startIndex]; } #endif } -static void i8VectorSumAVX2(const int8_t* plist, int32_t numOfRows, SAvgRes* pRes) { +static void i8VectorSumAVX2(const int8_t* plist, int32_t numOfRows, int32_t type, SAvgRes* pRes) { + const int32_t bitWidth = 256; + #if __AVX2__ // find the start position that are aligned to 32bytes address in memory - int32_t bitWidth = 16; - int32_t remainder = numOfRows % bitWidth; - int32_t rounds = numOfRows / bitWidth; + int32_t width = (bitWidth>>3u) / sizeof(int64_t); - const int8_t* p = plist; + int32_t remainder = numOfRows % width; + int32_t rounds = numOfRows / width; __m256i sum = _mm256_setzero_si256(); - for (int32_t i = 0; i < rounds; ++i) { - __m256i val = _mm256_lddqu_si256((__m256i*)p); -// __m256i extVal = _mm256_cvtepi8_epi64(val); - sum = _mm256_add_epi8(sum, val); - p += bitWidth; + if (type == TSDB_DATA_TYPE_TINYINT) { + const int8_t* p = plist; + + for (int32_t i = 0; i < rounds; ++i) { + __m128i val = _mm_lddqu_si128((__m128i*)p); + __m256i extVal = _mm256_cvtepi8_epi64(val); // only four items will be converted into __m256i + sum = _mm256_add_epi64(sum, extVal); + p += width; + } + } else { + const uint8_t* p = (const uint8_t*)plist; + + for(int32_t i = 0; i < rounds; ++i) { + __m128i val = _mm_lddqu_si128((__m128i*)p); + __m256i extVal = _mm256_cvtepu8_epi64(val); // only four items will be converted into __m256i + sum = _mm256_add_epi64(sum, extVal); + p += width; + } } // let sum up the final results - const int8_t* q = (const int8_t*)∑ + const int64_t* q = (const int64_t*)∑ pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; - int32_t startIndex = rounds * bitWidth; + int32_t startIndex = rounds * width; for (int32_t j = 0; j < remainder; ++j) { pRes->sum.isum += plist[j + startIndex]; } #endif } -static void i32VectorSumAVX2(const int32_t* plist, int32_t numOfRows, SAvgRes* pRes) { +static void i16VectorSumAVX2(const int16_t* plist, int32_t numOfRows, int32_t type, SAvgRes* pRes) { + const int32_t bitWidth = 256; + #if __AVX2__ // find the start position that are aligned to 32bytes address in memory - int32_t bitWidth = 8; - int32_t remainder = numOfRows % bitWidth; - int32_t rounds = numOfRows / bitWidth; + int32_t width = (bitWidth>>3u) / sizeof(int64_t); - const int32_t* p = plist; + int32_t remainder = numOfRows % width; + int32_t rounds = numOfRows / width; __m256i sum = _mm256_setzero_si256(); - for (int32_t i = 0; i < rounds; ++i) { - __m256i val = _mm256_lddqu_si256((__m256i*)p); - sum = _mm256_add_epi32(sum, val); - p += bitWidth; + + if (type == TSDB_DATA_TYPE_SMALLINT) { + const int16_t* p = plist; + + for (int32_t i = 0; i < rounds; ++i) { + __m128i val = _mm_lddqu_si128((__m128i*)p); + __m256i extVal = _mm256_cvtepi16_epi64(val); // only four items will be converted into __m256i + sum = _mm256_add_epi64(sum, extVal); + p += width; + } + } else { + const uint8_t* p = (const uint8_t*)plist; + + for(int32_t i = 0; i < rounds; ++i) { + __m128i val = _mm_lddqu_si128((__m128i*)p); + __m256i extVal = _mm256_cvtepu16_epi64(val); // only four items will be converted into __m256i + sum = _mm256_add_epi64(sum, extVal); + p += width; + } } // let sum up the final results - const int32_t* q = (const int32_t*)∑ - pRes->sum.isum += q[0] + q[1] + q[2] + q[3] + q[4] + q[5] + q[6] + q[7]; + const int64_t* q = (const int64_t*)∑ + pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; - int32_t startIndex = rounds * bitWidth; + int32_t startIndex = rounds * width; + for (int32_t j = 0; j < remainder; ++j) { + pRes->sum.isum += plist[j + startIndex]; + } +#endif +} + +static void i32VectorSumAVX2(const int32_t* plist, int32_t numOfRows, int32_t type, SAvgRes* pRes) { + const int32_t bitWidth = 256; + +#if __AVX2__ + // find the start position that are aligned to 32bytes address in memory + int32_t width = (bitWidth>>3u) / sizeof(int64_t); + + int32_t remainder = numOfRows % width; + int32_t rounds = numOfRows / width; + + __m256i sum = _mm256_setzero_si256(); + + if (type == TSDB_DATA_TYPE_INT) { + const int32_t* p = plist; + + for (int32_t i = 0; i < rounds; ++i) { + __m128i val = _mm_lddqu_si128((__m128i*)p); + __m256i extVal = _mm256_cvtepi32_epi64(val); // only four items will be converted into __m256i + sum = _mm256_add_epi64(sum, extVal); + p += width; + } + } else { + const uint32_t* p = (const uint32_t*)plist; + + for(int32_t i = 0; i < rounds; ++i) { + __m128i val = _mm_lddqu_si128((__m128i*)p); + __m256i extVal = _mm256_cvtepu32_epi64(val); // only four items will be converted into __m256i + sum = _mm256_add_epi64(sum, extVal); + p += width; + } + } + + // let sum up the final results + const int64_t* q = (const int64_t*)∑ + pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; + + int32_t startIndex = rounds * width; for (int32_t j = 0; j < remainder; ++j) { pRes->sum.isum += plist[j + startIndex]; } @@ -163,27 +242,30 @@ static void i32VectorSumAVX2(const int32_t* plist, int32_t numOfRows, SAvgRes* p } static void i64VectorSumAVX2(const int64_t* plist, int32_t numOfRows, SAvgRes* pRes) { + const int32_t bitWidth = 256; + #if __AVX2__ // find the start position that are aligned to 32bytes address in memory - int32_t bitWidth = 4; - int32_t remainder = numOfRows % bitWidth; - int32_t rounds = numOfRows / bitWidth; + int32_t width = (bitWidth>>3u) / sizeof(int64_t); - const int64_t* p = plist; + int32_t remainder = numOfRows % width; + int32_t rounds = numOfRows / width; __m256i sum = _mm256_setzero_si256(); + const int64_t* p = plist; + for (int32_t i = 0; i < rounds; ++i) { __m256i val = _mm256_lddqu_si256((__m256i*)p); sum = _mm256_add_epi64(sum, val); - p += bitWidth; + p += width; } // let sum up the final results const int64_t* q = (const int64_t*)∑ pRes->sum.isum += q[0] + q[1] + q[2] + q[3]; - int32_t startIndex = rounds * bitWidth; + int32_t startIndex = rounds * width; for (int32_t j = 0; j < remainder; ++j) { pRes->sum.isum += plist[j + startIndex]; } @@ -256,6 +338,163 @@ static int32_t calculateAvgBySMAInfo(SAvgRes* pRes, int32_t numOfRows, int32_t t return numOfElem; } +static int32_t doAddNumericVector(SColumnInfoData* pCol, int32_t type, SInputColumnInfoData *pInput, SAvgRes* pRes) { + int32_t start = pInput->startRowIndex; + int32_t numOfRows = pInput->numOfRows; + int32_t numOfElems = 0; + + switch (type) { + case TSDB_DATA_TYPE_TINYINT: { + int8_t* plist = (int8_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.isum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_SMALLINT: { + int16_t* plist = (int16_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.isum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_INT: { + int32_t* plist = (int32_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.isum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_BIGINT: { + int64_t* plist = (int64_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.isum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_UTINYINT: { + uint8_t* plist = (uint8_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.usum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_USMALLINT: { + uint16_t* plist = (uint16_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.usum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_UINT: { + uint32_t* plist = (uint32_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.usum += plist[i]; + } + + break; + } + + case TSDB_DATA_TYPE_UBIGINT: { + uint64_t* plist = (uint64_t*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.usum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_FLOAT: { + float* plist = (float*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.dsum += plist[i]; + } + break; + } + + case TSDB_DATA_TYPE_DOUBLE: { + double* plist = (double*)pCol->pData; + for (int32_t i = start; i < numOfRows + start; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + numOfElems += 1; + pRes->count += 1; + pRes->sum.dsum += plist[i]; + } + break; + } + + default: + break; + } + + return numOfElems; +} + int32_t avgFunction(SqlFunctionCtx* pCtx) { int32_t numOfElem = 0; const int32_t THRESHOLD_SIZE = 8; @@ -274,8 +513,7 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { int32_t numOfRows = pInput->numOfRows; if (IS_NULL_TYPE(type)) { - numOfElem = 0; - goto _avg_over; + goto _over; } if (pInput->colDataSMAIsSet) { // try to use SMA if available @@ -284,28 +522,31 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { numOfElem = pInput->numOfRows; pAvgRes->count += pInput->numOfRows; - bool simdAvaiable = tsAVXEnable && tsSIMDEnable && (numOfRows > THRESHOLD_SIZE); + bool simdAvailable = tsAVXEnable && tsSIMDEnable && (numOfRows > THRESHOLD_SIZE); switch(type) { + case TSDB_DATA_TYPE_UTINYINT: case TSDB_DATA_TYPE_TINYINT: { - const int8_t* plist = (const int8_t*) pCol->pData; + const int8_t* plist = (const int8_t*) &pCol->pData[start]; // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { - i8VectorSumAVX2(plist, numOfRows, pAvgRes); + if (simdAvailable) { + i8VectorSumAVX2(plist, numOfRows, type, pAvgRes); } else { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { - pAvgRes->sum.isum += plist[i]; + pAvgRes->sum.usum += plist[i]; } } break; } + + case TSDB_DATA_TYPE_USMALLINT: case TSDB_DATA_TYPE_SMALLINT: { - const double* plist = (const double*)pCol->pData; + const int16_t* plist = (const int16_t*)pCol->pData; // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { - doubleVectorSumAVX(plist, numOfRows, pAvgRes); + if (simdAvailable) { + i16VectorSumAVX2(plist, numOfRows, type, pAvgRes); } else { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { pAvgRes->sum.isum += plist[i]; @@ -313,12 +554,14 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { } break; } + + case TSDB_DATA_TYPE_UINT: case TSDB_DATA_TYPE_INT: { const int32_t* plist = (const int32_t*) pCol->pData; // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { - i32VectorSumAVX2(plist, numOfRows, pAvgRes); + if (simdAvailable) { + i32VectorSumAVX2(plist, numOfRows, type, pAvgRes); } else { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { pAvgRes->sum.isum += plist[i]; @@ -326,11 +569,13 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { } break; } + + case TSDB_DATA_TYPE_UBIGINT: case TSDB_DATA_TYPE_BIGINT: { const int64_t* plist = (const int64_t*) pCol->pData; // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { + if (simdAvailable) { i64VectorSumAVX2(plist, numOfRows, pAvgRes); } else { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { @@ -343,7 +588,7 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { const float* plist = (const float*) pCol->pData; // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { + if (simdAvailable) { floatVectorSumAVX(plist, numOfRows, pAvgRes); } else { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { @@ -353,10 +598,10 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { break; } case TSDB_DATA_TYPE_DOUBLE: { - const double* plist = (const double*) pCol->pData; + const double* plist = (const double*)pCol->pData; // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { + if (simdAvailable) { doubleVectorSumAVX(plist, numOfRows, pAvgRes); } else { for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { @@ -365,217 +610,14 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { } break; } - case TSDB_DATA_TYPE_UTINYINT: { - const double* plist = (const double*) pCol->pData; - - // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { - doubleVectorSumAVX(plist, numOfRows, pAvgRes); - } else { - for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { - pAvgRes->sum.usum += plist[i]; - } - } - break; - } - case TSDB_DATA_TYPE_USMALLINT: { - const double* plist = (const double*) pCol->pData; - - // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { - doubleVectorSumAVX(plist, numOfRows, pAvgRes); - } else { - for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { - pAvgRes->sum.usum += plist[i]; - } - } - break; - } - case TSDB_DATA_TYPE_UINT: { - const double* plist = (const double*) pCol->pData; - - // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { - doubleVectorSumAVX(plist, numOfRows, pAvgRes); - } else { - for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { - pAvgRes->sum.usum += plist[i]; - } - } - break; - } - case TSDB_DATA_TYPE_UBIGINT: { - const double* plist = (const double*) pCol->pData; - - // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (simdAvaiable) { - doubleVectorSumAVX(plist, numOfRows, pAvgRes); - } else { - for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { - pAvgRes->sum.usum += plist[i]; - } - } - break; - } default: ASSERT(0); } } else { - switch (type) { - case TSDB_DATA_TYPE_TINYINT: { - int8_t* plist = (int8_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_SMALLINT: { - int16_t* plist = (int16_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_INT: { - int32_t* plist = (int32_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_BIGINT: { - int64_t* plist = (int64_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.isum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_UTINYINT: { - uint8_t* plist = (uint8_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_USMALLINT: { - uint16_t* plist = (uint16_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_UINT: { - uint32_t* plist = (uint32_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - - break; - } - - case TSDB_DATA_TYPE_UBIGINT: { - uint64_t* plist = (uint64_t*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.usum += plist[i]; - } - break; - } - - case TSDB_DATA_TYPE_FLOAT: { -#if 1 - numOfElem = handleFloatCols(pCol, pInput, pAvgRes); -#else - float* plist = (float*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.dsum += plist[i]; - } -#endif - break; - } - - case TSDB_DATA_TYPE_DOUBLE: { - double* plist = (double*)pCol->pData; - for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElem += 1; - pAvgRes->count += 1; - pAvgRes->sum.dsum += plist[i]; - } - break; - } - - default: - break; - } + numOfElem = doAddNumericVector(pCol, type, pInput, pAvgRes); } -_avg_over: +_over: // data in the check operation are all null, not output SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); return TSDB_CODE_SUCCESS; diff --git a/source/util/src/thash.c b/source/util/src/thash.c index c3d4668e11..c542aa81a8 100644 --- a/source/util/src/thash.c +++ b/source/util/src/thash.c @@ -67,7 +67,7 @@ struct SHashObj { bool enableUpdate; // enable update SArray *pMemBlock; // memory block allocated for SHashEntry _hash_before_fn_t callbackFp; // function invoked before return the value to caller - int64_t compTimes; +// int64_t compTimes; }; /* @@ -147,7 +147,7 @@ static FORCE_INLINE SHashNode *doSearchInEntryList(SHashObj *pHashObj, SHashEntr uint32_t hashVal) { SHashNode *pNode = pe->next; while (pNode) { - atomic_add_fetch_64(&pHashObj->compTimes, 1); +// atomic_add_fetch_64(&pHashObj->compTimes, 1); if ((pNode->keyLen == keyLen) && ((*(pHashObj->equalFp))(GET_HASH_NODE_KEY(pNode), key, keyLen) == 0) && pNode->removed == 0) { assert(pNode->hashVal == hashVal); @@ -889,4 +889,4 @@ void *taosHashAcquire(SHashObj *pHashObj, const void *key, size_t keyLen) { void taosHashRelease(SHashObj *pHashObj, void *p) { taosHashCancelIterate(pHashObj, p); } -int64_t taosHashGetCompTimes(SHashObj *pHashObj) { return atomic_load_64(&pHashObj->compTimes); } +//int64_t taosHashGetCompTimes(SHashObj *pHashObj) { return atomic_load_64(&pHashObj->compTimes); } From ba5244d1b66a2ac301fe73e26cf7a10be6ad4939 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Tue, 15 Nov 2022 16:06:29 +0800 Subject: [PATCH 12/83] refactor: add tag filter results cache. --- source/dnode/vnode/src/inc/meta.h | 1 + source/dnode/vnode/src/meta/metaCache.c | 89 +++++++++++++++++++++++++ 2 files changed, 90 insertions(+) diff --git a/source/dnode/vnode/src/inc/meta.h b/source/dnode/vnode/src/inc/meta.h index 9e2fe4aaf0..17fdaa9815 100644 --- a/source/dnode/vnode/src/inc/meta.h +++ b/source/dnode/vnode/src/inc/meta.h @@ -70,6 +70,7 @@ int32_t metaCacheDrop(SMeta* pMeta, int64_t uid); int32_t metaStatsCacheUpsert(SMeta* pMeta, SMetaStbStats* pInfo); int32_t metaStatsCacheDrop(SMeta* pMeta, int64_t uid); int32_t metaStatsCacheGet(SMeta* pMeta, int64_t uid, SMetaStbStats* pInfo); +int32_t metaUidFilterCacheGet(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, LRUHandle** pHandle); struct SMeta { TdThreadRwlock lock; diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 356aa78c22..98bd9626e7 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -31,6 +31,12 @@ typedef struct SMetaStbStatsEntry { SMetaStbStats info; } SMetaStbStatsEntry; +typedef struct STagFilterResEntry { + uint64_t suid; // uid for super table + SList* pList; // the linked list of md5 digest, extracted from the serialized tag query condition + uint32_t qTimes;// queried times for current super table +} STagFilterResEntry; + struct SMetaCache { // child, normal, super, table entry cache struct SEntryCache { @@ -47,6 +53,10 @@ struct SMetaCache { } sStbStatsCache; // query cache + struct STagFilterResCache { + SHashObj* pTableEntry; + SLRUCache* pUidResCache; + } sTagFilterResCache; }; static void entryCacheClose(SMeta* pMeta) { @@ -388,3 +398,82 @@ int32_t metaStatsCacheGet(SMeta* pMeta, int64_t uid, SMetaStbStats* pInfo) { return code; } + +int32_t metaUidFilterCacheGet(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, LRUHandle** pHandle) { + // generate the composed key for LRU cache + char* p = taosMemoryMalloc(keyLen + sizeof(uint64_t)); + *(uint64_t*) p = suid; + memcpy(p + sizeof(suid), pKey, keyLen); + + int32_t len = keyLen + sizeof(uint64_t); + *pHandle = taosLRUCacheLookup(pMeta->pCache->sTagFilterResCache.pUidResCache, p, len); + if (*pHandle == NULL) { + taosMemoryFree(p); + return TSDB_CODE_SUCCESS; + } else { // do some book mark work after acquiring the filter result from cache + STagFilterResEntry* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); + ASSERT(pEntry != NULL); + + pEntry->qTimes += 1; + + // check if scanning all items are necessary or not + if (pEntry->qTimes > 5000 && TD_DLIST_NELES(pEntry->pList) > 10) { + SArray* pList = taosArrayInit(64, POINTER_BYTES); + + SListIter iter = {0}; + tdListInitIter(pEntry->pList, &iter, TD_LIST_FORWARD); + + SListNode* pNode = NULL; + while ((pNode = tdListNext(&iter)) != NULL) { + memcpy(p + sizeof(suid), pNode->data, keyLen); + + // check whether it is existed in LRU cache, and remove it from linked list if not. + void* pRes = taosLRUCacheLookup(pMeta->pCache->sTagFilterResCache.pUidResCache, p, len); + if (pRes == NULL) { // remove the item in the linked list + taosArrayPush(pList, &pNode); + } + } + + // remove the keys, of which query uid lists have been replaced already. + size_t s = taosArrayGetSize(pList); + for(int32_t i = 0; i < s; ++i) { + SListNode** p1 = taosArrayGet(pList, i); + tdListPopNode(pEntry->pList, *p1); + } + } + + taosMemoryFree(p); + } + + return TSDB_CODE_SUCCESS; +} + +// check both the payload size and selectivity ratio +int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, void* pPayload) { + + return TSDB_CODE_SUCCESS; +} + +// remove the lru cache that are expired due to the tags value update, or creating, or dropping, of child tables +int32_t metaUidCacheClear(SMeta* pMeta, uint64_t suid) { + STagFilterResEntry* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); + if (pEntry == NULL) { + return TSDB_CODE_SUCCESS; + } + + int32_t keyLen = sizeof(uint64_t) + 128; + char* p = taosMemoryMalloc(keyLen); + *(uint64_t*)p = pEntry->suid; + + SListIter iter = {0}; + tdListInitIter(pEntry->pList, &iter, TD_LIST_FORWARD); + + SListNode* pNode = NULL; + while ((pNode = tdListNext(&iter)) != NULL) { + memcpy(p + sizeof(suid), pNode->data, 128); + taosLRUCacheErase(pMeta->pCache->sTagFilterResCache.pUidResCache, p, keyLen); + } + + return TSDB_CODE_SUCCESS; +} + From 7d00b7a6c4f19b6a21562b95dfbca67e132be622 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 14:23:12 +0800 Subject: [PATCH 13/83] refactor: add tag filter cache. --- include/common/tglobal.h | 2 + include/util/tlist.h | 2 +- source/common/src/tglobal.c | 3 + source/dnode/vnode/inc/vnode.h | 3 + source/dnode/vnode/src/meta/metaCache.c | 99 +++++++++++++++++++------ source/libs/executor/src/executil.c | 50 ++++++++++--- source/util/src/tlist.c | 2 +- 7 files changed, 127 insertions(+), 34 deletions(-) diff --git a/include/common/tglobal.h b/include/common/tglobal.h index 681d1beb79..48886584d2 100644 --- a/include/common/tglobal.h +++ b/include/common/tglobal.h @@ -44,6 +44,8 @@ extern int32_t tsCompatibleModel; extern bool tsPrintAuth; extern int64_t tsTickPerMin[3]; extern int32_t tsCountAlwaysReturnValue; +extern float tsSelectivityRatio; +extern int32_t tsTagFilterResCacheSize; // queue & threads extern int32_t tsNumOfRpcThreads; diff --git a/include/util/tlist.h b/include/util/tlist.h index 1954bda145..3dbdb72f9e 100644 --- a/include/util/tlist.h +++ b/include/util/tlist.h @@ -225,7 +225,7 @@ void *tdListFree(SList *list); void tdListPrependNode(SList *list, SListNode *node); void tdListAppendNode(SList *list, SListNode *node); int32_t tdListPrepend(SList *list, void *data); -int32_t tdListAppend(SList *list, void *data); +int32_t tdListAppend(SList *list, const void *data); SListNode *tdListPopHead(SList *list); SListNode *tdListPopTail(SList *list); SListNode *tdListGetHead(SList *list); diff --git a/source/common/src/tglobal.c b/source/common/src/tglobal.c index 50b2c976fd..41be026a4c 100644 --- a/source/common/src/tglobal.c +++ b/source/common/src/tglobal.c @@ -119,6 +119,9 @@ int32_t tsMinIntervalTime = 1; // maximum memory allowed to be allocated for a single csv load (in MB) int32_t tsMaxMemUsedByInsert = 1024; +float tsSelectivityRatio = 1.0; +int32_t tsTagFilterResCacheSize = 4096; + // the maximum allowed query buffer size during query processing for each data node. // -1 no limit (default) // 0 no query allowed, queries are disabled diff --git a/source/dnode/vnode/inc/vnode.h b/source/dnode/vnode/inc/vnode.h index 0b58959822..d09724ffc4 100644 --- a/source/dnode/vnode/inc/vnode.h +++ b/source/dnode/vnode/inc/vnode.h @@ -108,6 +108,9 @@ int metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName); int metaGetTableUidByName(void *meta, char *tbName, uint64_t *uid); int metaGetTableTypeByName(void *meta, char *tbName, ETableType *tbType); bool metaIsTableExist(SMeta *pMeta, tb_uid_t uid); +int32_t metaGetCachedTableUidList(SMeta *pMeta, tb_uid_t suid, const uint8_t *key, int32_t keyLen, SArray *pList, bool* acquired); +int32_t metaUidFilterCachePut(SMeta *pMeta, uint64_t suid, const void *pKey, int32_t keyLen, void *pPayload, + int32_t payloadLen, double selectivityRatio); typedef struct SMetaFltParam { tb_uid_t suid; diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 98bd9626e7..e448c6e7c6 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -33,7 +33,7 @@ typedef struct SMetaStbStatsEntry { typedef struct STagFilterResEntry { uint64_t suid; // uid for super table - SList* pList; // the linked list of md5 digest, extracted from the serialized tag query condition + SList list; // the linked list of md5 digest, extracted from the serialized tag query condition uint32_t qTimes;// queried times for current super table } STagFilterResEntry; @@ -56,6 +56,7 @@ struct SMetaCache { struct STagFilterResCache { SHashObj* pTableEntry; SLRUCache* pUidResCache; + uint64_t keyBuf[3]; } sTagFilterResCache; }; @@ -119,9 +120,19 @@ int32_t metaCacheOpen(SMeta* pMeta) { goto _err2; } - pMeta->pCache = pCache; + pCache->sTagFilterResCache.pUidResCache = taosLRUCacheInit(5*1024*1024, -1, 0.5); + if (pCache->sTagFilterResCache.pUidResCache == NULL) { + code = TSDB_CODE_OUT_OF_MEMORY; + goto _err2; + } -_exit: + pCache->sTagFilterResCache.pTableEntry = taosHashInit(1024, taosGetDefaultHashFunction(TSDB_DATA_TYPE_VARCHAR), false, HASH_NO_LOCK); + if (pCache->sTagFilterResCache.pTableEntry == NULL) { + code = TSDB_CODE_OUT_OF_MEMORY; + goto _err2; + } + + pMeta->pCache = pCache; return code; _err2: @@ -129,7 +140,6 @@ _err2: _err: taosMemoryFree(pCache); - metaError("vgId:%d, meta open cache failed since %s", TD_VID(pMeta->pVnode), tstrerror(code)); return code; } @@ -138,6 +148,11 @@ void metaCacheClose(SMeta* pMeta) { if (pMeta->pCache) { entryCacheClose(pMeta); statsCacheClose(pMeta); + + taosHashCleanup(pMeta->pCache->sTagFilterResCache.pTableEntry); + taosLRUCacheCleanup(pMeta->pCache->sTagFilterResCache.pUidResCache); + + taosMemoryFree(pMeta->pCache->sTagFilterResCache.keyBuf); taosMemoryFree(pMeta->pCache); pMeta->pCache = NULL; } @@ -399,38 +414,48 @@ int32_t metaStatsCacheGet(SMeta* pMeta, int64_t uid, SMetaStbStats* pInfo) { return code; } -int32_t metaUidFilterCacheGet(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, LRUHandle** pHandle) { +int32_t metaGetCachedTableUidList(SMeta* pMeta, tb_uid_t suid, const uint8_t* pKey, int32_t keyLen, SArray* pList1, bool* acquireRes) { + uint64_t* pBuf = pMeta->pCache->sTagFilterResCache.keyBuf; + // generate the composed key for LRU cache - char* p = taosMemoryMalloc(keyLen + sizeof(uint64_t)); - *(uint64_t*) p = suid; - memcpy(p + sizeof(suid), pKey, keyLen); + SLRUCache* pCache = pMeta->pCache->sTagFilterResCache.pUidResCache; + + pBuf[0] = suid; + memcpy(&pBuf[1], pKey, keyLen); int32_t len = keyLen + sizeof(uint64_t); - *pHandle = taosLRUCacheLookup(pMeta->pCache->sTagFilterResCache.pUidResCache, p, len); - if (*pHandle == NULL) { - taosMemoryFree(p); + LRUHandle *pHandle = taosLRUCacheLookup(pCache, pBuf, len); + if (pHandle == NULL) { + *acquireRes = 0; return TSDB_CODE_SUCCESS; } else { // do some book mark work after acquiring the filter result from cache STagFilterResEntry* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); ASSERT(pEntry != NULL); + *acquireRes = 1; + + const char* p = taosLRUCacheValue(pMeta->pCache->sTagFilterResCache.pUidResCache, pHandle); + int32_t size = *(int32_t*) p; + taosArrayAddBatch(pList1, p + sizeof(int32_t), size); pEntry->qTimes += 1; // check if scanning all items are necessary or not - if (pEntry->qTimes > 5000 && TD_DLIST_NELES(pEntry->pList) > 10) { + if (pEntry->qTimes >= 5000 && TD_DLIST_NELES(&pEntry->list) > 10) { SArray* pList = taosArrayInit(64, POINTER_BYTES); SListIter iter = {0}; - tdListInitIter(pEntry->pList, &iter, TD_LIST_FORWARD); + tdListInitIter(&pEntry->list, &iter, TD_LIST_FORWARD); SListNode* pNode = NULL; while ((pNode = tdListNext(&iter)) != NULL) { - memcpy(p + sizeof(suid), pNode->data, keyLen); + memcpy(pBuf + sizeof(suid), pNode->data, keyLen); // check whether it is existed in LRU cache, and remove it from linked list if not. - void* pRes = taosLRUCacheLookup(pMeta->pCache->sTagFilterResCache.pUidResCache, p, len); + LRUHandle* pRes = taosLRUCacheLookup(pCache, pBuf, len); if (pRes == NULL) { // remove the item in the linked list taosArrayPush(pList, &pNode); + } else { + taosLRUCacheRelease(pCache, pRes, false); } } @@ -438,19 +463,49 @@ int32_t metaUidFilterCacheGet(SMeta* pMeta, uint64_t suid, const void* pKey, int size_t s = taosArrayGetSize(pList); for(int32_t i = 0; i < s; ++i) { SListNode** p1 = taosArrayGet(pList, i); - tdListPopNode(pEntry->pList, *p1); + tdListPopNode(&pEntry->list, *p1); } - } - taosMemoryFree(p); + pEntry->qTimes = 0; // reset the query times + } } return TSDB_CODE_SUCCESS; } // check both the payload size and selectivity ratio -int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, void* pPayload) { +int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, void* pPayload, int32_t payloadLen, double selectivityRatio) { + if (selectivityRatio > tsSelectivityRatio) { + return TSDB_CODE_SUCCESS; + } + if (payloadLen > tsTagFilterResCacheSize) { + return TSDB_CODE_SUCCESS; + } + + SLRUCache* pCache = pMeta->pCache->sTagFilterResCache.pUidResCache; + SHashObj* pTableEntry = pMeta->pCache->sTagFilterResCache.pTableEntry; + + void* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); + if (pEntry == NULL) { + STagFilterResEntry* p = taosMemoryMalloc(sizeof(STagFilterResEntry)); + p->qTimes = 0; + tdListInit(&p->list, keyLen); + taosHashPut(pTableEntry, &suid, sizeof(uint64_t), pEntry, POINTER_BYTES); + + pEntry = &p; + } + + tdListAppend(&(*(STagFilterResEntry**)pEntry)->list, pKey); + + uint64_t* pBuf = pMeta->pCache->sTagFilterResCache.keyBuf; + pBuf[0] = suid; + + memcpy(&pBuf[1], pKey, keyLen); + ASSERT(sizeof(uint64_t) + keyLen == 24); + + // add to cache. + taosLRUCacheInsert(pCache, pBuf, sizeof(uint64_t) + keyLen, pPayload, payloadLen, NULL, NULL, TAOS_LRU_PRIORITY_LOW); return TSDB_CODE_SUCCESS; } @@ -466,7 +521,7 @@ int32_t metaUidCacheClear(SMeta* pMeta, uint64_t suid) { *(uint64_t*)p = pEntry->suid; SListIter iter = {0}; - tdListInitIter(pEntry->pList, &iter, TD_LIST_FORWARD); + tdListInitIter(&pEntry->list, &iter, TD_LIST_FORWARD); SListNode* pNode = NULL; while ((pNode = tdListNext(&iter)) != NULL) { @@ -474,6 +529,8 @@ int32_t metaUidCacheClear(SMeta* pMeta, uint64_t suid) { taosLRUCacheErase(pMeta->pCache->sTagFilterResCache.pUidResCache, p, keyLen); } + pEntry->qTimes = 0; + tdListEmpty(&pEntry->list); + return TSDB_CODE_SUCCESS; } - diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 44390ca2e5..8604fd4db9 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -973,19 +973,47 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, SArray* res = taosArrayInit(8, sizeof(uint64_t)); if (pScanNode->tableType == TSDB_SUPER_TABLE) { - if (pTagIndexCond) { - SIndexMetaArg metaArg = { - .metaEx = metaHandle, .idx = tsdbGetIdx(metaHandle), .ivtIdx = tsdbGetIvtIdx(metaHandle), .suid = tableUid}; + // try to retrieve the result from meta cache + // generate the cache key + T_MD5_CTX context = {0}; - // int64_t stt = taosGetTimestampUs(); - SIdxFltStatus status = SFLT_NOT_INDEX; - code = doFilterTag(pTagIndexCond, &metaArg, res, &status); - if (code != 0 || status == SFLT_NOT_INDEX) { - qError("failed to get tableIds from index, reason:%s, suid:%" PRIu64, tstrerror(code), tableUid); - code = TDB_CODE_SUCCESS; + if (pTagIndexCond) { + char* payload = NULL; + int32_t len = 0; + nodesNodeToMsg(pTagCond, &payload, &len); + + tMD5Init(&context); + tMD5Update(&context, (uint8_t*)payload, (uint32_t)len); + tMD5Final(&context); + } + + bool acquired = false; + metaGetCachedTableUidList(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), res, &acquired); + if (!acquired) { + // failed to find the result in the cache, let try to calculate the results + if (pTagIndexCond) { + SIndexMetaArg metaArg = { + .metaEx = metaHandle, .idx = tsdbGetIdx(metaHandle), .ivtIdx = tsdbGetIvtIdx(metaHandle), .suid = tableUid}; + + SIdxFltStatus status = SFLT_NOT_INDEX; + code = doFilterTag(pTagIndexCond, &metaArg, res, &status); + if (code != 0 || status == SFLT_NOT_INDEX) { + qError("failed to get tableIds from index, reason:%s, suid:%" PRIu64, tstrerror(code), tableUid); + code = TDB_CODE_SUCCESS; + } + } else if (!pTagCond) { + vnodeGetCtbIdList(pVnode, pScanNode->suid, res); } - } else if (!pTagCond) { - vnodeGetCtbIdList(pVnode, pScanNode->suid, res); + + // let's add the filter results into meta-cache + size_t numOfTables = taosArrayGetSize(res); + size_t size = numOfTables * sizeof(uint64_t) + sizeof(int32_t); + char* pPayload = taosMemoryMalloc(size); + *(int32_t*)pPayload = numOfTables; + memcpy(pPayload + sizeof(int32_t), taosArrayGet(res, 0), numOfTables * sizeof(uint64_t)); + + metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, + size, 1); } } else { // Create one table group. if (metaIsTableExist(metaHandle, tableUid)) { diff --git a/source/util/src/tlist.c b/source/util/src/tlist.c index b1c0188051..1b12ea0cdd 100644 --- a/source/util/src/tlist.c +++ b/source/util/src/tlist.c @@ -60,7 +60,7 @@ int32_t tdListPrepend(SList *list, void *data) { return 0; } -int32_t tdListAppend(SList *list, void *data) { +int32_t tdListAppend(SList *list, const void *data) { SListNode *node = (SListNode *)taosMemoryCalloc(1, sizeof(SListNode) + list->eleSize); if (node == NULL) return -1; From 2320c0d2bd2863a0a08845893545e6e30e9db3d6 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 14:40:39 +0800 Subject: [PATCH 14/83] refactor: update some logs. --- source/libs/executor/src/executil.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 8604fd4db9..17e41b3a08 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1942,13 +1942,15 @@ int32_t createScanTableListInfo(SScanPhysiNode* pScanNode, SNodeList* pGroupTags return code; } + int32_t numOfTables = taosArrayGetSize(pTableListInfo->pTableList); ASSERT(pTableListInfo->numOfOuputGroups == 1); int64_t st1 = taosGetTimestampUs(); pTaskInfo->cost.extractListTime = (st1 - st) / 1000.0; - qDebug("extract queried table list completed, elapsed time:%.2f ms %s", pTaskInfo->cost.extractListTime, idStr); + qDebug("extract queried table list completed, %d tables, elapsed time:%.2f ms %s", numOfTables, + pTaskInfo->cost.extractListTime, idStr); - if (taosArrayGetSize(pTableListInfo->pTableList) == 0) { + if (numOfTables == 0) { qDebug("no table qualified for query, %s" PRIx64, idStr); return TSDB_CODE_SUCCESS; } From 47a7664e2c8fe67aee5ed371703e78a96e803fe9 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 14:52:38 +0800 Subject: [PATCH 15/83] refactor: add some logs. --- source/common/src/tglobal.c | 2 +- source/dnode/vnode/src/meta/metaCache.c | 9 +++++++++ 2 files changed, 10 insertions(+), 1 deletion(-) diff --git a/source/common/src/tglobal.c b/source/common/src/tglobal.c index 41be026a4c..0db3e99863 100644 --- a/source/common/src/tglobal.c +++ b/source/common/src/tglobal.c @@ -120,7 +120,7 @@ int32_t tsMinIntervalTime = 1; int32_t tsMaxMemUsedByInsert = 1024; float tsSelectivityRatio = 1.0; -int32_t tsTagFilterResCacheSize = 4096; +int32_t tsTagFilterResCacheSize = 1024*10; // the maximum allowed query buffer size during query processing for each data node. // -1 no limit (default) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index e448c6e7c6..17c99e34bf 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -476,10 +476,16 @@ int32_t metaGetCachedTableUidList(SMeta* pMeta, tb_uid_t suid, const uint8_t* pK // check both the payload size and selectivity ratio int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, void* pPayload, int32_t payloadLen, double selectivityRatio) { if (selectivityRatio > tsSelectivityRatio) { + metaDebug("vgId:%d, suid:%" PRIu64 + " failed to add to uid list cache, due to selectivity ratio %.2f less than threshold %.2f", + TD_VID(pMeta->pVnode), suid, selectivityRatio, tsSelectivityRatio); return TSDB_CODE_SUCCESS; } if (payloadLen > tsTagFilterResCacheSize) { + metaDebug("vgId:%d, suid:%" PRIu64 + " failed to add to uid list cache, due to payload length %d greater than threshold %d", + TD_VID(pMeta->pVnode), suid, payloadLen, tsTagFilterResCacheSize); return TSDB_CODE_SUCCESS; } @@ -506,6 +512,9 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int // add to cache. taosLRUCacheInsert(pCache, pBuf, sizeof(uint64_t) + keyLen, pPayload, payloadLen, NULL, NULL, TAOS_LRU_PRIORITY_LOW); + metaDebug("vgId:%d, suid:%"PRIu64" list cache added into cache, total:%d, tables:%d", TD_VID(pMeta->pVnode), + suid, (int32_t) taosLRUCacheGetUsage(pCache), taosHashGetSize(pTableEntry)); + return TSDB_CODE_SUCCESS; } From 23da7bc5e98b26708cae61f5b246babcb3d79169 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 15:13:38 +0800 Subject: [PATCH 16/83] refactor: add some logs. --- source/libs/executor/src/executil.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 17e41b3a08..28afa7a41e 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1014,6 +1014,8 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); + } else { + qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)) } } else { // Create one table group. if (metaIsTableExist(metaHandle, tableUid)) { From 82f22e407d04013d686219122a5836d49d856975 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 15:13:54 +0800 Subject: [PATCH 17/83] refactor: do some internal refactor. --- source/libs/executor/src/executil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 28afa7a41e..7de93780b3 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1015,7 +1015,7 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); } else { - qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)) + qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)); } } else { // Create one table group. if (metaIsTableExist(metaHandle, tableUid)) { From fade0507fa103313e00ff745749cd4015d4434cd Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 15:17:13 +0800 Subject: [PATCH 18/83] refactor: do some internal refactor. --- source/libs/executor/src/executil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 7de93780b3..68f4319bf4 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1064,7 +1064,7 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, return TSDB_CODE_OUT_OF_MEMORY; } - qDebug("tagfilter get uid:%" PRIu64 "", info.uid); + qTrace("tagfilter get uid:%" PRIu64 "", info.uid); } taosArrayDestroy(res); From 2ee5fa87d0441f35f646a04ba0801c7257682486 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 15:19:15 +0800 Subject: [PATCH 19/83] refactor: do some internal refactor. --- source/libs/executor/src/executil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 68f4319bf4..cebed9c3f4 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -988,7 +988,7 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, } bool acquired = false; - metaGetCachedTableUidList(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), res, &acquired); +// metaGetCachedTableUidList(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), res, &acquired); if (!acquired) { // failed to find the result in the cache, let try to calculate the results if (pTagIndexCond) { From 819235dd3b735bbaac6be2620b2f838c5a27ca6d Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 15:21:33 +0800 Subject: [PATCH 20/83] refactor: do some internal refactor. --- source/libs/executor/src/executil.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index cebed9c3f4..b774b48726 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1006,6 +1006,7 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, } // let's add the filter results into meta-cache +#if 0 size_t numOfTables = taosArrayGetSize(res); size_t size = numOfTables * sizeof(uint64_t) + sizeof(int32_t); char* pPayload = taosMemoryMalloc(size); @@ -1014,6 +1015,8 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); +#endif + } else { qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)); } From 42f8abcc0e0f2ec7281345ea797002a2400fb8f8 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 15:24:30 +0800 Subject: [PATCH 21/83] refactor: do some internal refactor. --- source/libs/executor/src/executil.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index b774b48726..68f4319bf4 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -988,7 +988,7 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, } bool acquired = false; -// metaGetCachedTableUidList(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), res, &acquired); + metaGetCachedTableUidList(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), res, &acquired); if (!acquired) { // failed to find the result in the cache, let try to calculate the results if (pTagIndexCond) { @@ -1006,7 +1006,6 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, } // let's add the filter results into meta-cache -#if 0 size_t numOfTables = taosArrayGetSize(res); size_t size = numOfTables * sizeof(uint64_t) + sizeof(int32_t); char* pPayload = taosMemoryMalloc(size); @@ -1015,8 +1014,6 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); -#endif - } else { qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)); } From a36c24eb1ceac64c8b9cfdfb812345d726544c72 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 16 Nov 2022 15:47:24 +0800 Subject: [PATCH 22/83] fix(query): remove invalid free. --- source/dnode/vnode/src/meta/metaCache.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 17c99e34bf..22fc49efac 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -151,8 +151,6 @@ void metaCacheClose(SMeta* pMeta) { taosHashCleanup(pMeta->pCache->sTagFilterResCache.pTableEntry); taosLRUCacheCleanup(pMeta->pCache->sTagFilterResCache.pUidResCache); - - taosMemoryFree(pMeta->pCache->sTagFilterResCache.keyBuf); taosMemoryFree(pMeta->pCache); pMeta->pCache = NULL; } From ec7ad45b2b8d2ffb8b2f7aa931db8c5a47b85b18 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 17 Nov 2022 17:37:45 +0800 Subject: [PATCH 23/83] fix(query): fix a typo. --- source/libs/function/src/detail/tavgfunction.c | 6 ++---- source/libs/qworker/src/qwUtil.c | 4 ++-- 2 files changed, 4 insertions(+), 6 deletions(-) diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index d7ef73b08e..744927d6c8 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -31,7 +31,7 @@ do { \ T* plist = (T*)pCol->pData; \ for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { \ - if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { \ + if (colDataIsNull_f(pCol->nullbitmap, i)) { \ continue; \ } \ \ @@ -661,8 +661,6 @@ int32_t avgInvertFunction(SqlFunctionCtx* pCtx) { // Only the pre-computing information loaded and actual data does not loaded SInputColumnInfoData* pInput = &pCtx->input; - int32_t type = pInput->pData[0]->info.type; - SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); // computing based on the true data block @@ -671,7 +669,7 @@ int32_t avgInvertFunction(SqlFunctionCtx* pCtx) { int32_t start = pInput->startRowIndex; int32_t numOfRows = pInput->numOfRows; - switch (type) { + switch (pCol->info.type) { case TSDB_DATA_TYPE_TINYINT: { LIST_AVG_N(pAvgRes->sum.isum, int8_t); break; diff --git a/source/libs/qworker/src/qwUtil.c b/source/libs/qworker/src/qwUtil.c index e9ded9b269..80a0a6e0ae 100644 --- a/source/libs/qworker/src/qwUtil.c +++ b/source/libs/qworker/src/qwUtil.c @@ -275,7 +275,7 @@ void qwFreeTaskHandle(qTaskInfo_t *taskHandle) { qTaskInfo_t otaskHandle = atomic_load_ptr(taskHandle); if (otaskHandle && atomic_val_compare_exchange_ptr(taskHandle, otaskHandle, NULL)) { qDestroyTask(otaskHandle); - qDebug("task handle destryed"); + qDebug("task handle destroyed"); } } @@ -306,7 +306,7 @@ void qwFreeTaskCtx(SQWTaskCtx *ctx) { if (ctx->sinkHandle) { dsDestroyDataSinker(ctx->sinkHandle); ctx->sinkHandle = NULL; - qDebug("sink handle destryed"); + qDebug("sink handle destroyed"); } } From f26a492c678a6ff79fc33bf27a31374184878592 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 18 Nov 2022 14:00:27 +0800 Subject: [PATCH 24/83] fix(query): check result size before store it. --- source/libs/executor/src/executil.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 68f4319bf4..6ee2dce59b 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1010,7 +1010,10 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, size_t size = numOfTables * sizeof(uint64_t) + sizeof(int32_t); char* pPayload = taosMemoryMalloc(size); *(int32_t*)pPayload = numOfTables; - memcpy(pPayload + sizeof(int32_t), taosArrayGet(res, 0), numOfTables * sizeof(uint64_t)); + + if (numOfTables > 0) { + memcpy(pPayload + sizeof(int32_t), taosArrayGet(res, 0), numOfTables * sizeof(uint64_t)); + } metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); From 3ca38c41a704a4ad325caff1ef5837b809cbb95d Mon Sep 17 00:00:00 2001 From: Minglei Jin Date: Fri, 18 Nov 2022 14:38:44 +0800 Subject: [PATCH 25/83] meta: use meta cache to get uid's version instead of fetching from tdb --- source/dnode/vnode/src/meta/metaQuery.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaQuery.c b/source/dnode/vnode/src/meta/metaQuery.c index 32dd427d09..89eb0525ea 100644 --- a/source/dnode/vnode/src/meta/metaQuery.c +++ b/source/dnode/vnode/src/meta/metaQuery.c @@ -152,7 +152,8 @@ bool metaIsTableExist(SMeta *pMeta, tb_uid_t uid) { } int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) { - SMeta *pMeta = pReader->pMeta; + SMeta *pMeta = pReader->pMeta; + /* int64_t version1; // query uid.idx @@ -163,6 +164,14 @@ int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) { version1 = ((SUidIdxVal *)pReader->pBuf)[0].version; return metaGetTableEntryByVersion(pReader, version1, uid); + */ + SMetaInfo info; + if (metaGetInfo(pMeta, uid, &info) == TSDB_CODE_NOT_FOUND) { + terrno = TSDB_CODE_PAR_TABLE_NOT_EXIST; + return -1; + } + + return metaGetTableEntryByVersion(pReader, info.version, uid); } int metaGetTableEntryByName(SMetaReader *pReader, const char *name) { @@ -1126,7 +1135,7 @@ int32_t metaFilterTableName(SMeta *pMeta, SMetaFltParam *param, SArray *pUids) { valid = tdbTbcGet(pCursor->pCur, (const void **)pEntryKey, &nEntryKey, (const void **)&pEntryVal, &nEntryVal); if (valid < 0) break; - char *pTableKey = (char *)pEntryKey; + char *pTableKey = (char *)pEntryKey; cmp = (*param->filterFunc)(pTableKey, pName, pCursor->type); if (cmp == 0) { tb_uid_t tuid = *(tb_uid_t *)pEntryVal; From 71de2f056b92afb9b61c3c869d22561a525500a3 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 18 Nov 2022 14:53:39 +0800 Subject: [PATCH 26/83] fix(query): fix invalid write/read. --- source/dnode/vnode/src/meta/metaCache.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 22fc49efac..8647250191 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -427,7 +427,7 @@ int32_t metaGetCachedTableUidList(SMeta* pMeta, tb_uid_t suid, const uint8_t* pK *acquireRes = 0; return TSDB_CODE_SUCCESS; } else { // do some book mark work after acquiring the filter result from cache - STagFilterResEntry* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); + STagFilterResEntry** pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); ASSERT(pEntry != NULL); *acquireRes = 1; @@ -435,14 +435,14 @@ int32_t metaGetCachedTableUidList(SMeta* pMeta, tb_uid_t suid, const uint8_t* pK int32_t size = *(int32_t*) p; taosArrayAddBatch(pList1, p + sizeof(int32_t), size); - pEntry->qTimes += 1; + (*pEntry)->qTimes += 1; // check if scanning all items are necessary or not - if (pEntry->qTimes >= 5000 && TD_DLIST_NELES(&pEntry->list) > 10) { + if ((*pEntry)->qTimes >= 5000 && TD_DLIST_NELES(&(*pEntry)->list) > 10) { SArray* pList = taosArrayInit(64, POINTER_BYTES); SListIter iter = {0}; - tdListInitIter(&pEntry->list, &iter, TD_LIST_FORWARD); + tdListInitIter(&(*pEntry)->list, &iter, TD_LIST_FORWARD); SListNode* pNode = NULL; while ((pNode = tdListNext(&iter)) != NULL) { @@ -461,10 +461,10 @@ int32_t metaGetCachedTableUidList(SMeta* pMeta, tb_uid_t suid, const uint8_t* pK size_t s = taosArrayGetSize(pList); for(int32_t i = 0; i < s; ++i) { SListNode** p1 = taosArrayGet(pList, i); - tdListPopNode(&pEntry->list, *p1); + tdListPopNode(&(*pEntry)->list, *p1); } - pEntry->qTimes = 0; // reset the query times + (*pEntry)->qTimes = 0; // reset the query times } } From 35b9dc6b6cc0c0829ae092ac37311caad1a37ade Mon Sep 17 00:00:00 2001 From: Minglei Jin Date: Fri, 18 Nov 2022 14:38:44 +0800 Subject: [PATCH 27/83] meta: use meta cache to get uid's version instead of fetching from tdb --- source/dnode/vnode/src/inc/vnodeInt.h | 2 +- source/dnode/vnode/src/meta/metaQuery.c | 30 +++++++++++++++++++--- source/dnode/vnode/src/meta/metaTable.c | 2 +- source/dnode/vnode/src/tsdb/tsdbMemTable.c | 6 ++--- 4 files changed, 31 insertions(+), 9 deletions(-) diff --git a/source/dnode/vnode/src/inc/vnodeInt.h b/source/dnode/vnode/src/inc/vnodeInt.h index ac9fabf052..f229b3b127 100644 --- a/source/dnode/vnode/src/inc/vnodeInt.h +++ b/source/dnode/vnode/src/inc/vnodeInt.h @@ -142,7 +142,7 @@ typedef struct SMetaInfo { int64_t version; int32_t skmVer; } SMetaInfo; -int32_t metaGetInfo(SMeta* pMeta, int64_t uid, SMetaInfo* pInfo); +int32_t metaGetInfo(SMeta* pMeta, int64_t uid, SMetaInfo* pInfo, SMetaReader* pReader); typedef struct { int64_t uid; diff --git a/source/dnode/vnode/src/meta/metaQuery.c b/source/dnode/vnode/src/meta/metaQuery.c index 32dd427d09..f58b97eeb8 100644 --- a/source/dnode/vnode/src/meta/metaQuery.c +++ b/source/dnode/vnode/src/meta/metaQuery.c @@ -152,7 +152,8 @@ bool metaIsTableExist(SMeta *pMeta, tb_uid_t uid) { } int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) { - SMeta *pMeta = pReader->pMeta; + SMeta *pMeta = pReader->pMeta; + /* int64_t version1; // query uid.idx @@ -163,6 +164,15 @@ int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) { version1 = ((SUidIdxVal *)pReader->pBuf)[0].version; return metaGetTableEntryByVersion(pReader, version1, uid); + */ + + SMetaInfo info; + if (metaGetInfo(pMeta, uid, &info, pReader) == TSDB_CODE_NOT_FOUND) { + terrno = TSDB_CODE_PAR_TABLE_NOT_EXIST; + return -1; + } + + return metaGetTableEntryByVersion(pReader, info.version, uid); } int metaGetTableEntryByName(SMetaReader *pReader, const char *name) { @@ -614,7 +624,7 @@ int32_t metaGetTbTSchemaEx(SMeta *pMeta, tb_uid_t suid, tb_uid_t uid, int32_t sv SSkmDbKey skmDbKey; if (sver <= 0) { SMetaInfo info; - if (metaGetInfo(pMeta, suid ? suid : uid, &info) == 0) { + if (metaGetInfo(pMeta, suid ? suid : uid, &info, NULL) == 0) { sver = info.skmVer; } else { TBC *pSkmDbC = NULL; @@ -1126,7 +1136,7 @@ int32_t metaFilterTableName(SMeta *pMeta, SMetaFltParam *param, SArray *pUids) { valid = tdbTbcGet(pCursor->pCur, (const void **)pEntryKey, &nEntryKey, (const void **)&pEntryVal, &nEntryVal); if (valid < 0) break; - char *pTableKey = (char *)pEntryKey; + char *pTableKey = (char *)pEntryKey; cmp = (*param->filterFunc)(pTableKey, pName, pCursor->type); if (cmp == 0) { tb_uid_t tuid = *(tb_uid_t *)pEntryVal; @@ -1379,10 +1389,11 @@ int32_t metaGetTableTags(SMeta *pMeta, uint64_t suid, SArray *uidList, SHashObj int32_t metaCacheGet(SMeta *pMeta, int64_t uid, SMetaInfo *pInfo); -int32_t metaGetInfo(SMeta *pMeta, int64_t uid, SMetaInfo *pInfo) { +int32_t metaGetInfo(SMeta *pMeta, int64_t uid, SMetaInfo *pInfo, SMetaReader *pReader) { int32_t code = 0; void *pData = NULL; int nData = 0; + int lock = 0; metaRLock(pMeta); @@ -1407,11 +1418,22 @@ int32_t metaGetInfo(SMeta *pMeta, int64_t uid, SMetaInfo *pInfo) { pInfo->version = ((SUidIdxVal *)pData)->version; pInfo->skmVer = ((SUidIdxVal *)pData)->skmVer; + if (pReader != NULL) { + lock = !(pReader->flags & META_READER_NOLOCK); + if (lock) { + metaULock(pReader->pMeta); + // metaReaderReleaseLock(pReader); + } + } // upsert the cache metaWLock(pMeta); metaCacheUpsert(pMeta, pInfo); metaULock(pMeta); + if (lock) { + metaRLock(pReader->pMeta); + } + _exit: tdbFree(pData); return code; diff --git a/source/dnode/vnode/src/meta/metaTable.c b/source/dnode/vnode/src/meta/metaTable.c index 6dadce80ca..bb0cfe183d 100644 --- a/source/dnode/vnode/src/meta/metaTable.c +++ b/source/dnode/vnode/src/meta/metaTable.c @@ -207,7 +207,7 @@ int metaCreateSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) { tb_uid_t uid = *(tb_uid_t *)pData; tdbFree(pData); SMetaInfo info; - metaGetInfo(pMeta, uid, &info); + metaGetInfo(pMeta, uid, &info, NULL); if (info.uid == info.suid) { return 0; } else { diff --git a/source/dnode/vnode/src/tsdb/tsdbMemTable.c b/source/dnode/vnode/src/tsdb/tsdbMemTable.c index c663e2b526..ca3c283d05 100644 --- a/source/dnode/vnode/src/tsdb/tsdbMemTable.c +++ b/source/dnode/vnode/src/tsdb/tsdbMemTable.c @@ -104,7 +104,7 @@ int32_t tsdbInsertTableData(STsdb *pTsdb, int64_t version, SSubmitMsgIter *pMsgI tb_uid_t uid = pMsgIter->uid; SMetaInfo info; - code = metaGetInfo(pTsdb->pVnode->pMeta, uid, &info); + code = metaGetInfo(pTsdb->pVnode->pMeta, uid, &info, NULL); if (code) { code = TSDB_CODE_TDB_TABLE_NOT_EXIST; goto _err; @@ -114,7 +114,7 @@ int32_t tsdbInsertTableData(STsdb *pTsdb, int64_t version, SSubmitMsgIter *pMsgI goto _err; } if (info.suid) { - metaGetInfo(pTsdb->pVnode->pMeta, info.suid, &info); + metaGetInfo(pTsdb->pVnode->pMeta, info.suid, &info, NULL); } if (pMsgIter->sversion != info.skmVer) { tsdbError("vgId:%d, req sver:%d, skmVer:%d suid:%" PRId64 " uid:%" PRId64, TD_VID(pTsdb->pVnode), @@ -153,7 +153,7 @@ int32_t tsdbDeleteTableData(STsdb *pTsdb, int64_t version, tb_uid_t suid, tb_uid // check if table exists SMetaInfo info; - code = metaGetInfo(pTsdb->pVnode->pMeta, uid, &info); + code = metaGetInfo(pTsdb->pVnode->pMeta, uid, &info, NULL); if (code) { code = TSDB_CODE_TDB_TABLE_NOT_EXIST; goto _err; From d3452f6630538068be2541a4c38e8a932327e40a Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 18 Nov 2022 17:38:31 +0800 Subject: [PATCH 28/83] fix(query): fix error. --- source/dnode/vnode/src/meta/metaCache.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 8647250191..05b6209c56 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -495,9 +495,9 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int STagFilterResEntry* p = taosMemoryMalloc(sizeof(STagFilterResEntry)); p->qTimes = 0; tdListInit(&p->list, keyLen); - taosHashPut(pTableEntry, &suid, sizeof(uint64_t), pEntry, POINTER_BYTES); pEntry = &p; + taosHashPut(pTableEntry, &suid, sizeof(uint64_t), pEntry, POINTER_BYTES); } tdListAppend(&(*(STagFilterResEntry**)pEntry)->list, pKey); From 1e25eac4c7257f00006fcaf2f866039b2faf5f7c Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sun, 20 Nov 2022 23:11:12 +0800 Subject: [PATCH 29/83] refactor: do some internal refactor. --- source/libs/function/src/detail/tminmax.c | 577 +++++++++++++++------- 1 file changed, 396 insertions(+), 181 deletions(-) diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index d239315e0e..ed297e2b66 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -20,68 +20,59 @@ #include "tglobal.h" static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool isMinFunc) { - int32_t v = 0; + int32_t v = 0; + const int32_t bitWidth = 256; + const int32_t* p = pData; + + int32_t width = (bitWidth>>3u) / sizeof(int32_t); + int32_t remain = numOfRows % width; + int32_t rounds = numOfRows / width; #if __AVX2__ - int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); - int32_t bitWidth = 8; - - int32_t remain = (numOfRows - startElem) % bitWidth; - int32_t rounds = (numOfRows - startElem) / bitWidth; - const int32_t* p = &pData[startElem]; - __m256i next; __m256i initialVal = _mm256_loadu_si256((__m256i*)p); - p += bitWidth; + p += width; if (!isMinFunc) { // max function for (int32_t i = 0; i < rounds; ++i) { next = _mm256_lddqu_si256((__m256i*)p); initialVal = _mm256_max_epi32(initialVal, next); - p += bitWidth; + p += width; } // let sum up the final results const int32_t* q = (const int32_t*)&initialVal; - v = TMAX(q[0], q[1]); - v = TMAX(v, q[2]); - v = TMAX(v, q[3]); - v = TMAX(v, q[4]); - v = TMAX(v, q[5]); - v = TMAX(v, q[6]); - v = TMAX(v, q[7]); + for (int32_t k = 1; k < width; ++k) { + v = TMAX(v, q[k]); + } // calculate the front and the reminder items in array list - startElem += rounds * bitWidth; + int32_t start = rounds * width; for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + startElem]) { - v = p[j + startElem]; + if (v < p[j + start]) { + v = p[j + start]; } } } else { // min function for (int32_t i = 0; i < rounds; ++i) { next = _mm256_lddqu_si256((__m256i*)p); initialVal = _mm256_min_epi32(initialVal, next); - p += bitWidth; + p += width; } // let sum up the final results const int32_t* q = (const int32_t*)&initialVal; - v = TMIN(q[0], q[1]); - v = TMIN(v, q[2]); - v = TMIN(v, q[3]); - v = TMIN(v, q[4]); - v = TMIN(v, q[5]); - v = TMIN(v, q[6]); - v = TMIN(v, q[7]); + for (int32_t k = 1; k < width; ++k) { + v = TMIN(v, q[k]); + } // calculate the front and the remainder items in array list - startElem += rounds * bitWidth; + int32_t start = rounds * width; for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + startElem]) { - v = p[j + startElem]; + if (v > p[j + start]) { + v = p[j + start]; } } } @@ -92,69 +83,59 @@ static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool is static float floatVectorCmpAVX(const float* pData, int32_t numOfRows, bool isMinFunc) { float v = 0; + const int32_t bitWidth = 256; + const float* p = pData; + + int32_t width = (bitWidth>>3u) / sizeof(float); + int32_t remain = numOfRows % width; + int32_t rounds = numOfRows / width; #if __AVX__ - int32_t startElem = 0;//((uint64_t)plist) & ((1<<8u)-1); - int32_t i = 0; - - int32_t bitWidth = 8; - - int32_t remain = (numOfRows - startElem) % bitWidth; - int32_t rounds = (numOfRows - startElem) / bitWidth; - const float* p = &pData[startElem]; __m256 next; __m256 initialVal = _mm256_loadu_ps(p); - p += bitWidth; + p += width; if (!isMinFunc) { // max function - for (; i < rounds; ++i) { + for (int32_t i = 1; i < rounds; ++i) { next = _mm256_loadu_ps(p); initialVal = _mm256_max_ps(initialVal, next); - p += bitWidth; + p += width; } // let sum up the final results const float* q = (const float*)&initialVal; - v = TMAX(q[0], q[1]); - v = TMAX(v, q[2]); - v = TMAX(v, q[3]); - v = TMAX(v, q[4]); - v = TMAX(v, q[5]); - v = TMAX(v, q[6]); - v = TMAX(v, q[7]); + for (int32_t k = 1; k < width; ++k) { + v = TMAX(v, q[k]); + } // calculate the front and the reminder items in array list - startElem += rounds * bitWidth; + int32_t start = rounds * width; for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + startElem]) { - v = p[j + startElem]; + if (v < p[j + width]) { + v = p[j + width]; } } } else { // min function - for (; i < rounds; ++i) { + for (int32_t i = 1; i < rounds; ++i) { next = _mm256_loadu_ps(p); initialVal = _mm256_min_ps(initialVal, next); - p += bitWidth; + p += width; } // let sum up the final results const float* q = (const float*)&initialVal; - v = TMIN(q[0], q[1]); - v = TMIN(v, q[2]); - v = TMIN(v, q[3]); - v = TMIN(v, q[4]); - v = TMIN(v, q[5]); - v = TMIN(v, q[6]); - v = TMIN(v, q[7]); + for (int32_t k = 1; k < width; ++k) { + v = TMIN(v, q[k]); + } // calculate the front and the reminder items in array list - startElem += rounds * bitWidth; + int32_t start = rounds * bitWidth; for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + startElem]) { - v = p[j + startElem]; + if (v > p[j + start]) { + v = p[j + start]; } } } @@ -163,6 +144,195 @@ static float floatVectorCmpAVX(const float* pData, int32_t numOfRows, bool isMin return v; } +static int8_t i8VectorCmpAVX2(const int8_t* pData, int32_t numOfRows, bool isMinFunc) { + int8_t v = 0; + const int32_t bitWidth = 256; + const int8_t* p = pData; + + int32_t width = (bitWidth>>3u) / sizeof(int8_t); + int32_t remain = numOfRows % width; + int32_t rounds = numOfRows / width; + +#if __AVX2__ + __m256i next; + __m256i initialVal = _mm256_loadu_si256((__m256i*)p); + p += width; + + if (!isMinFunc) { // max function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initialVal = _mm256_max_epi8(initialVal, next); + p += width; + } + + // let sum up the final results + const int8_t* q = (const int8_t*)&initialVal; + v = TMAX(q[0], q[1]); + for (int32_t k = 1; k < width; ++k) { + v = TMAX(v, q[k]); + } + + // calculate the front and the reminder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + start]) { + v = p[j + start]; + } + } + } else { // min function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initialVal = _mm256_min_epi8(initialVal, next); + p += width; + } + + // let sum up the final results + const int8_t* q = (const int8_t*)&initialVal; + + v = TMIN(q[0], q[1]); + for(int32_t k = 1; k < width; ++k) { + v = TMIN(v, q[k]); + } + + // calculate the front and the remainder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + start]) { + v = p[j + start]; + } + } + } +#endif + + return v; +} + +static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool isMinFunc) { + int16_t v = 0; + const int32_t bitWidth = 256; + const int16_t* p = pData; + + int32_t width = (bitWidth>>3u) / sizeof(int16_t); + int32_t remain = numOfRows % width; + int32_t rounds = numOfRows / width; + +#if __AVX2__ + __m256i next; + __m256i initialVal = _mm256_loadu_si256((__m256i*)p); + p += width; + + if (!isMinFunc) { // max function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initialVal = _mm256_max_epi16(initialVal, next); + p += width; + } + + // let sum up the final results + const int16_t* q = (const int16_t*)&initialVal; + + v = TMAX(q[0], q[1]); + for(int32_t k = 1; k < width; ++k) { + v = TMAX(v, q[k]); + } + + // calculate the front and the reminder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + start]) { + v = p[j + start]; + } + } + } else { // min function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initialVal = _mm256_min_epi16(initialVal, next); + p += width; + } + + // let sum up the final results + const int16_t* q = (const int16_t*)&initialVal; + + v = TMIN(q[0], q[1]); + for(int32_t k = 1; k < width; ++k) { + v = TMIN(v, q[k]); + } + + // calculate the front and the remainder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + start]) { + v = p[j + start]; + } + } + } +#endif + + return v; +} + +//static int64_t i64VectorCmpAVX2(const int64_t* pData, int32_t numOfRows, bool isMinFunc) { +// int64_t v = 0; +// const int32_t bitWidth = 256; +// const int64_t* p = pData; +// +// int32_t width = (bitWidth>>3u) / sizeof(int64_t); +// int32_t remain = numOfRows % width; +// int32_t rounds = numOfRows / width; +// +//#if __AVX2__ +// __m256i next; +// __m256i initialVal = _mm256_loadu_si256((__m256i*)p); +// p += width; +// +// if (!isMinFunc) { // max function +// for (int32_t i = 0; i < rounds; ++i) { +// next = _mm256_lddqu_si256((__m256i*)p); +// initialVal = _mm256_max_epi64(initialVal, next); +// p += width; +// } +// +// // let sum up the final results +// const int64_t* q = (const int64_t*)&initialVal; +// v = TMAX(q[0], q[1]); +// for(int32_t k = 1; k < width; ++k) { +// v = TMAX(v, q[k]); +// } +// +// // calculate the front and the reminder items in array list +// int32_t start = rounds * width; +// for (int32_t j = 0; j < remain; ++j) { +// if (v < p[j + start]) { +// v = p[j + start]; +// } +// } +// } else { // min function +// for (int32_t i = 0; i < rounds; ++i) { +// next = _mm256_lddqu_si256((__m256i*)p); +// initialVal = _mm256_min_epi64(initialVal, next); +// p += width; +// } +// +// // let sum up the final results +// const int64_t* q = (const int64_t*)&initialVal; +// v = TMIN(q[0], q[1]); +// for(int32_t k = 1; k < width; ++k) { +// v = TMIN(v, q[k]); +// } +// +// // calculate the front and the remainder items in array list +// int32_t start = rounds * width; +// for (int32_t j = 0; j < remain; ++j) { +// if (v > p[j + start]) { +// v = p[j + start]; +// } +// } +// } +//#endif +// +// return v; +//} + static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { int32_t* pData = (int32_t*)pCol->pData; @@ -170,56 +340,56 @@ static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numO int32_t numOfElems = 0; if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { - if (isMinFunc) { // min - for (int32_t i = start; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - - numOfElems += 1; - } - } else { // max function - for (int32_t i = start; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - - numOfElems += 1; + int32_t i = start; + while (i < (start + numOfRows)) { + if (!colDataIsNull_f(pCol->nullbitmap, i)) { + break; } + i += 1; } - } else { // not has null value + + if ((i < (start + numOfRows)) && (!pBuf->assign)) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + numOfElems += 1; + } + + if (isMinFunc) { // min + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } else { // max function + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } + } else { // not has null value // AVX2 version to speedup the loop if (tsAVX2Enable && tsSIMDEnable) { *val = i32VectorCmpAVX2(pData, numOfRows, isMinFunc); @@ -257,56 +427,55 @@ static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numO int32_t numOfElems = 0; if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { + int32_t i = start; + while (i < (start + numOfRows)) { + if (!colDataIsNull_f(pCol->nullbitmap, i)) { + break; + } + i += 1; + } + + if ((i < (start + numOfRows)) && (!pBuf->assign)) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + numOfElems += 1; + } + if (isMinFunc) { // min - for (int32_t i = start; i < start + numOfRows; ++i) { + for (; i < start + numOfRows; ++i) { if (colDataIsNull_f(pCol->nullbitmap, i)) { continue; } - if (!pBuf->assign) { + if (*val > pData[i]) { *val = pData[i]; if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); } } - numOfElems += 1; } } else { // max function - for (int32_t i = start; i < start + numOfRows; ++i) { + for (; i < start + numOfRows; ++i) { if (colDataIsNull_f(pCol->nullbitmap, i)) { continue; } - if (!pBuf->assign) { + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { *val = pData[i]; if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); } } - numOfElems += 1; } } - } else { // not has null value + } else { // not has null value // AVX version to speedup the loop if (tsAVXEnable && tsSIMDEnable) { *val = (double) floatVectorCmpAVX(pData, numOfRows, isMinFunc); @@ -337,6 +506,93 @@ static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numO return numOfElems; } +static int32_t handleInt8Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, + SMinmaxResInfo* pBuf, bool isMinFunc) { + int8_t* pData = (int8_t*)pCol->pData; + int8_t* val = (int8_t*)&pBuf->v; + + int32_t numOfElems = 0; + if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { + int32_t i = start; + while (i < (start + numOfRows)) { + if (!colDataIsNull_f(pCol->nullbitmap, i)) { + break; + } + i += 1; + } + + if ((i < (start + numOfRows)) && (!pBuf->assign)) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + numOfElems += 1; + } + + if (isMinFunc) { // min + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } else { // max function + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } + } else { // not has null value + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + *val = i8VectorCmpAVX2(pData, numOfRows, isMinFunc); + } else { + if (!pBuf->assign) { + *val = pData[0]; + pBuf->assign = true; + } + + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val > pData[i]) { + *val = pData[i]; + } + } + } else { // max + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val < pData[i]) { + *val = pData[i]; + } + } + } + } + + numOfElems = numOfRows; + } + + return numOfElems; +} + static int32_t findRowIndex(int32_t start, int32_t num, SColumnInfoData* pCol, const char* tval) { // the data is loaded, not only the block SMA value for (int32_t i = start; i < num + start; ++i) { @@ -463,42 +719,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { if (IS_SIGNED_NUMERIC_TYPE(type) || type == TSDB_DATA_TYPE_BOOL) { if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) { - int8_t* pData = (int8_t*)pCol->pData; - int8_t* val = (int8_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } + numOfElems = handleInt8Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); } else if (type == TSDB_DATA_TYPE_SMALLINT) { int16_t* pData = (int16_t*)pCol->pData; int16_t* val = (int16_t*)&pBuf->v; @@ -537,9 +758,6 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { numOfElems += 1; } } else if (type == TSDB_DATA_TYPE_INT) { - int32_t* pData = (int32_t*)pCol->pData; - int32_t* val = (int32_t*)&pBuf->v; - numOfElems = handleInt32Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); #if 0 for (int32_t i = start; i < start + numOfRows; ++i) { @@ -803,9 +1021,6 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { numOfElems += 1; } } else if (type == TSDB_DATA_TYPE_FLOAT) { - float* pData = (float*)pCol->pData; - float* val = (float*)&pBuf->v; - numOfElems = handleFloatCol(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); #if 0 for (int32_t i = start; i < start + numOfRows; ++i) { From 8fee089aaf30f30b36de81409d057dbf48a4f4bf Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Mon, 21 Nov 2022 22:48:25 +0800 Subject: [PATCH 30/83] refactor: do some internal refactor. --- source/dnode/vnode/src/tsdb/tsdbRead.c | 1 + 1 file changed, 1 insertion(+) diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index 6ea270e5f4..e919593f90 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -693,6 +693,7 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN tMapDataReset(&pScanInfo->mapData); tsdbReadDataBlk(pReader->pFileReader, pBlockIdx, &pScanInfo->mapData); + taosArrayEnsureCap(pScanInfo->pBlockList, pScanInfo->mapData.nItem); sizeInDisk += pScanInfo->mapData.nData; for (int32_t j = 0; j < pScanInfo->mapData.nItem; ++j) { From 56edf57c121f97efc6bccacfe1dfc21c679ad4cd Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Tue, 22 Nov 2022 09:52:53 +0800 Subject: [PATCH 31/83] refactor: do some internal refactor. --- source/dnode/vnode/src/tsdb/tsdbRead.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index e919593f90..04f2a14bce 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -38,7 +38,7 @@ typedef struct { typedef struct SBlockIndex { int32_t ordinalIndex; int64_t inFileOffset; - STimeWindow window; + STimeWindow window; // todo replace it with overlap flag. } SBlockIndex; typedef struct STableBlockScanInfo { @@ -1440,6 +1440,7 @@ static int32_t setFileBlockActiveInBlockIter(SDataBlockIter* pBlockIter, int32_t return TSDB_CODE_SUCCESS; } +// todo: this attribute could be acquired during extractin the global ordered block list. static bool overlapWithNeighborBlock(SDataBlk* pBlock, SBlockIndex* pNeighborBlockIndex, int32_t order) { // it is the last block in current file, no chance to overlap with neighbor blocks. if (ASCENDING_TRAVERSE(order)) { From 390709e3c856817f595e101e7c7d3b31a9423554 Mon Sep 17 00:00:00 2001 From: Hongze Cheng Date: Tue, 22 Nov 2022 10:10:23 +0800 Subject: [PATCH 32/83] add interface --- include/util/tarray.h | 9 +++++++++ source/dnode/vnode/src/inc/tsdb.h | 2 ++ source/dnode/vnode/src/tsdb/tsdbUtil.c | 24 ++++++++++++++++++++++++ source/util/src/tarray.c | 11 +++++++++++ 4 files changed, 46 insertions(+) diff --git a/include/util/tarray.h b/include/util/tarray.h index e95568197b..0632db3103 100644 --- a/include/util/tarray.h +++ b/include/util/tarray.h @@ -104,6 +104,15 @@ static FORCE_INLINE void* taosArrayPush(SArray* pArray, const void* pData) { return taosArrayAddBatch(pArray, pData, 1); } +/** + * @brief reserve the capacity of the array + * + * @param pArray + * @param num + * @return void* the start position of the reserved memory + */ +void* taosArrayReserve(SArray* pArray, int32_t num); + /** * * @param pArray diff --git a/source/dnode/vnode/src/inc/tsdb.h b/source/dnode/vnode/src/inc/tsdb.h index a5257b32c0..6552a76590 100644 --- a/source/dnode/vnode/src/inc/tsdb.h +++ b/source/dnode/vnode/src/inc/tsdb.h @@ -186,6 +186,8 @@ int32_t tMapDataSearch(SMapData *pMapData, void *pSearchItem, int32_t (*tGetItem int32_t (*tItemCmprFn)(const void *, const void *), void *pItem); int32_t tPutMapData(uint8_t *p, SMapData *pMapData); int32_t tGetMapData(uint8_t *p, SMapData *pMapData); +int32_t tMapDataToArray(SMapData *pMapData, int32_t itemSize, int32_t (*tGetItemFn)(uint8_t *, void *), + SArray **ppArray); // other int32_t tsdbKeyFid(TSKEY key, int32_t minutes, int8_t precision); void tsdbFidKeyRange(int32_t fid, int32_t minutes, int8_t precision, TSKEY *minKey, TSKEY *maxKey); diff --git a/source/dnode/vnode/src/tsdb/tsdbUtil.c b/source/dnode/vnode/src/tsdb/tsdbUtil.c index 52b74aea3f..0902a9bd73 100644 --- a/source/dnode/vnode/src/tsdb/tsdbUtil.c +++ b/source/dnode/vnode/src/tsdb/tsdbUtil.c @@ -101,6 +101,30 @@ void tMapDataGetItemByIdx(SMapData *pMapData, int32_t idx, void *pItem, int32_t tGetItemFn(pMapData->pData + pMapData->aOffset[idx], pItem); } +int32_t tMapDataToArray(SMapData *pMapData, int32_t itemSize, int32_t (*tGetItemFn)(uint8_t *, void *), + SArray **ppArray) { + int32_t code = 0; + + SArray *pArray = taosArrayInit(pMapData->nItem, itemSize); + if (pArray == NULL) { + code = TSDB_CODE_TDB_OUT_OF_MEMORY; + goto _exit; + } + + for (int32_t i = 0; i < pMapData->nItem; i++) { + tMapDataGetItemByIdx(pMapData, i, taosArrayReserve(pArray, 1), tGetItemFn); + } + +_exit: + if (code) { + *ppArray = NULL; + if (pArray) taosArrayDestroy(pArray); + } else { + *ppArray = pArray; + } + return code; +} + int32_t tPutMapData(uint8_t *p, SMapData *pMapData) { int32_t n = 0; diff --git a/source/util/src/tarray.c b/source/util/src/tarray.c index 95065972a3..5703d8f8f4 100644 --- a/source/util/src/tarray.c +++ b/source/util/src/tarray.c @@ -181,6 +181,17 @@ void* taosArrayAddAll(SArray* pArray, const SArray* pInput) { } } +void* taosArrayReserve(SArray* pArray, int32_t num) { + if (taosArrayEnsureCap(pArray, pArray->size + num) != 0) { + return NULL; + } + + void* dst = TARRAY_GET_ELEM(pArray, pArray->size); + pArray->size += num; + + return dst; +} + void* taosArrayPop(SArray* pArray) { assert(pArray != NULL); From 0af77e6ae23c69a87f776d6a8efaa69bb4b2fa72 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Tue, 22 Nov 2022 12:43:00 +0800 Subject: [PATCH 33/83] refactor: do some internal refactor. --- source/dnode/vnode/src/tsdb/tsdbRead.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index 04f2a14bce..64cdba5648 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -695,33 +695,38 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN tsdbReadDataBlk(pReader->pFileReader, pBlockIdx, &pScanInfo->mapData); taosArrayEnsureCap(pScanInfo->pBlockList, pScanInfo->mapData.nItem); + SArray* pList = NULL; + tMapDataToArray(&pScanInfo->mapData, sizeof(SDataBlk), tGetDataBlk, &pList); sizeInDisk += pScanInfo->mapData.nData; for (int32_t j = 0; j < pScanInfo->mapData.nItem; ++j) { - SDataBlk block = {0}; - tMapDataGetItemByIdx(&pScanInfo->mapData, j, &block, tGetDataBlk); +// SDataBlk block = {0}; +// tMapDataGetItemByIdx(&pScanInfo->mapData, j, &block, tGetDataBlk); + SDataBlk* pBlock = taosArrayGet(pList, j); // 1. time range check - if (block.minKey.ts > pReader->window.ekey || block.maxKey.ts < pReader->window.skey) { + if (pBlock->minKey.ts > pReader->window.ekey || pBlock->maxKey.ts < pReader->window.skey) { continue; } // 2. version range check - if (block.minVer > pReader->verRange.maxVer || block.maxVer < pReader->verRange.minVer) { + if (pBlock->minVer > pReader->verRange.maxVer || pBlock->maxVer < pReader->verRange.minVer) { continue; } - SBlockIndex bIndex = {.ordinalIndex = j, .inFileOffset = block.aSubBlock->offset}; - bIndex.window = (STimeWindow){.skey = block.minKey.ts, .ekey = block.maxKey.ts}; + SBlockIndex bIndex = {.ordinalIndex = j, .inFileOffset = pBlock->aSubBlock->offset}; + bIndex.window = (STimeWindow){.skey = pBlock->minKey.ts, .ekey = pBlock->maxKey.ts}; void* p = taosArrayPush(pScanInfo->pBlockList, &bIndex); if (p == NULL) { tMapDataClear(&pScanInfo->mapData); + taosArrayDestroy(pList); return TSDB_CODE_OUT_OF_MEMORY; } pBlockNum->numOfBlocks += 1; } + taosArrayDestroy(pList); if (pScanInfo->pBlockList != NULL && taosArrayGetSize(pScanInfo->pBlockList) > 0) { numOfQTable += 1; } From 93ad9e9d3d05383c7f353cc0a5b8abbf73219ce2 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Tue, 22 Nov 2022 14:29:30 +0800 Subject: [PATCH 34/83] refactor: do some internal refactor. --- source/dnode/vnode/src/tsdb/tsdbRead.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index 64cdba5648..5b9ab2dac0 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -695,13 +695,14 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN tsdbReadDataBlk(pReader->pFileReader, pBlockIdx, &pScanInfo->mapData); taosArrayEnsureCap(pScanInfo->pBlockList, pScanInfo->mapData.nItem); - SArray* pList = NULL; - tMapDataToArray(&pScanInfo->mapData, sizeof(SDataBlk), tGetDataBlk, &pList); + SDataBlk* p = taosMemoryMalloc(sizeof(SDataBlk) * pScanInfo->mapData.nItem); + for (int32_t k = 0; k < pScanInfo->mapData.nItem; k++) { + tMapDataGetItemByIdx(&pScanInfo->mapData, i, &p[k], tGetDataBlk); + } + sizeInDisk += pScanInfo->mapData.nData; for (int32_t j = 0; j < pScanInfo->mapData.nItem; ++j) { -// SDataBlk block = {0}; -// tMapDataGetItemByIdx(&pScanInfo->mapData, j, &block, tGetDataBlk); - SDataBlk* pBlock = taosArrayGet(pList, j); + SDataBlk* pBlock = &p[j]; // 1. time range check if (pBlock->minKey.ts > pReader->window.ekey || pBlock->maxKey.ts < pReader->window.skey) { @@ -719,14 +720,14 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN void* p = taosArrayPush(pScanInfo->pBlockList, &bIndex); if (p == NULL) { tMapDataClear(&pScanInfo->mapData); - taosArrayDestroy(pList); + taosMemoryFree(p); return TSDB_CODE_OUT_OF_MEMORY; } pBlockNum->numOfBlocks += 1; } - taosArrayDestroy(pList); + taosMemoryFree(p); if (pScanInfo->pBlockList != NULL && taosArrayGetSize(pScanInfo->pBlockList) > 0) { numOfQTable += 1; } @@ -742,6 +743,7 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN numOfTables, pBlockNum->numOfBlocks, numOfQTable, pBlockNum->numOfLastFiles, sizeInDisk / 1000.0, el, pReader->idStr); + pReader->cost.numOfBlocks += total; pReader->cost.headFileLoadTime += el; From 4d206fd3b29cc2af9a755185df999e3090806034 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Tue, 22 Nov 2022 14:59:17 +0800 Subject: [PATCH 35/83] refactor: do some internal refactor. --- source/dnode/vnode/src/tsdb/tsdbRead.c | 29 ++++++++++++++------------ 1 file changed, 16 insertions(+), 13 deletions(-) diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index 5b9ab2dac0..d64c20ae9e 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -695,39 +695,42 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN tsdbReadDataBlk(pReader->pFileReader, pBlockIdx, &pScanInfo->mapData); taosArrayEnsureCap(pScanInfo->pBlockList, pScanInfo->mapData.nItem); - SDataBlk* p = taosMemoryMalloc(sizeof(SDataBlk) * pScanInfo->mapData.nItem); - for (int32_t k = 0; k < pScanInfo->mapData.nItem; k++) { - tMapDataGetItemByIdx(&pScanInfo->mapData, i, &p[k], tGetDataBlk); - } +// SDataBlk* p = taosMemoryMalloc(sizeof(SDataBlk) * pScanInfo->mapData.nItem); +// for (int32_t k = 0; k < pScanInfo->mapData.nItem; k++) { +// tMapDataGetItemByIdx(&pScanInfo->mapData, k, &p[k], tGetDataBlk); +// } sizeInDisk += pScanInfo->mapData.nData; + SDataBlk block = {0}; + for (int32_t j = 0; j < pScanInfo->mapData.nItem; ++j) { - SDataBlk* pBlock = &p[j]; + tGetDataBlk(pScanInfo->mapData.pData + pScanInfo->mapData.aOffset[j], &block); +// SDataBlk* pBlock = &p[j]; // 1. time range check - if (pBlock->minKey.ts > pReader->window.ekey || pBlock->maxKey.ts < pReader->window.skey) { + if (block.minKey.ts > pReader->window.ekey || block.maxKey.ts < pReader->window.skey) { continue; } // 2. version range check - if (pBlock->minVer > pReader->verRange.maxVer || pBlock->maxVer < pReader->verRange.minVer) { + if (block.minVer > pReader->verRange.maxVer || block.maxVer < pReader->verRange.minVer) { continue; } - SBlockIndex bIndex = {.ordinalIndex = j, .inFileOffset = pBlock->aSubBlock->offset}; - bIndex.window = (STimeWindow){.skey = pBlock->minKey.ts, .ekey = pBlock->maxKey.ts}; + SBlockIndex bIndex = {.ordinalIndex = j, .inFileOffset = block.aSubBlock->offset}; + bIndex.window = (STimeWindow){.skey = block.minKey.ts, .ekey = block.maxKey.ts}; - void* p = taosArrayPush(pScanInfo->pBlockList, &bIndex); - if (p == NULL) { + void* p1 = taosArrayPush(pScanInfo->pBlockList, &bIndex); + if (p1 == NULL) { tMapDataClear(&pScanInfo->mapData); - taosMemoryFree(p); +// taosMemoryFree(p); return TSDB_CODE_OUT_OF_MEMORY; } pBlockNum->numOfBlocks += 1; } - taosMemoryFree(p); +// taosMemoryFree(p); if (pScanInfo->pBlockList != NULL && taosArrayGetSize(pScanInfo->pBlockList) > 0) { numOfQTable += 1; } From b70a616a2d22e8d07fc906bd805658b2be508e87 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Tue, 22 Nov 2022 15:27:33 +0800 Subject: [PATCH 36/83] refactor: do some internal refactor. --- source/dnode/vnode/src/tsdb/tsdbRead.c | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index d64c20ae9e..c19e39e7e1 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -695,17 +695,11 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN tsdbReadDataBlk(pReader->pFileReader, pBlockIdx, &pScanInfo->mapData); taosArrayEnsureCap(pScanInfo->pBlockList, pScanInfo->mapData.nItem); -// SDataBlk* p = taosMemoryMalloc(sizeof(SDataBlk) * pScanInfo->mapData.nItem); -// for (int32_t k = 0; k < pScanInfo->mapData.nItem; k++) { -// tMapDataGetItemByIdx(&pScanInfo->mapData, k, &p[k], tGetDataBlk); -// } - sizeInDisk += pScanInfo->mapData.nData; - SDataBlk block = {0}; + SDataBlk block = {0}; for (int32_t j = 0; j < pScanInfo->mapData.nItem; ++j) { tGetDataBlk(pScanInfo->mapData.pData + pScanInfo->mapData.aOffset[j], &block); -// SDataBlk* pBlock = &p[j]; // 1. time range check if (block.minKey.ts > pReader->window.ekey || block.maxKey.ts < pReader->window.skey) { @@ -723,14 +717,12 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, SBlockN void* p1 = taosArrayPush(pScanInfo->pBlockList, &bIndex); if (p1 == NULL) { tMapDataClear(&pScanInfo->mapData); -// taosMemoryFree(p); return TSDB_CODE_OUT_OF_MEMORY; } pBlockNum->numOfBlocks += 1; } -// taosMemoryFree(p); if (pScanInfo->pBlockList != NULL && taosArrayGetSize(pScanInfo->pBlockList) > 0) { numOfQTable += 1; } From 611ab8b39407b5b7beababce8f4ef481bc433eeb Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 23 Nov 2022 21:52:51 +0800 Subject: [PATCH 37/83] refactor: make sure the memory is aligned to 32 bytes. --- include/os/osMemory.h | 1 + source/common/src/tdatablock.c | 13 ++++- source/libs/function/src/detail/tminmax.c | 61 ----------------------- source/os/src/osMemory.c | 8 +++ 4 files changed, 21 insertions(+), 62 deletions(-) diff --git a/include/os/osMemory.h b/include/os/osMemory.h index 14d53a7a06..4681ff6674 100644 --- a/include/os/osMemory.h +++ b/include/os/osMemory.h @@ -37,6 +37,7 @@ void taosMemoryFree(void *ptr); int64_t taosMemorySize(void *ptr); void taosPrintBackTrace(); void taosMemoryTrim(int32_t size); +void *taosMemoryMallocAlign(uint32_t alignment, int64_t size); #define taosMemoryFreeClear(ptr) \ do { \ diff --git a/source/common/src/tdatablock.c b/source/common/src/tdatablock.c index f5dba35440..c79910978a 100644 --- a/source/common/src/tdatablock.c +++ b/source/common/src/tdatablock.c @@ -19,6 +19,8 @@ #include "tlog.h" #include "tname.h" +#define MALLOC_ALIGN_BYTES 32 + int32_t colDataGetLength(const SColumnInfoData* pColumnInfoData, int32_t numOfRows) { ASSERT(pColumnInfoData != NULL); if (IS_VAR_DATA_TYPE(pColumnInfoData->info.type)) { @@ -1163,6 +1165,7 @@ static int32_t doEnsureCapacity(SColumnInfoData* pColumn, const SDataBlockInfo* pColumn->varmeta.offset = (int32_t*)tmp; memset(&pColumn->varmeta.offset[existedRows], 0, sizeof(int32_t) * (numOfRows - existedRows)); } else { + // prepare for the null bitmap char* tmp = taosMemoryRealloc(pColumn->nullbitmap, BitmapLen(numOfRows)); if (tmp == NULL) { return TSDB_CODE_OUT_OF_MEMORY; @@ -1173,11 +1176,19 @@ static int32_t doEnsureCapacity(SColumnInfoData* pColumn, const SDataBlockInfo* memset(&pColumn->nullbitmap[oldLen], 0, BitmapLen(numOfRows) - oldLen); ASSERT(pColumn->info.bytes); - tmp = taosMemoryRealloc(pColumn->pData, numOfRows * pColumn->info.bytes); + + // make sure the allocated memory is MALLOC_ALIGN_BYTES aligned + tmp = taosMemoryMallocAlign(MALLOC_ALIGN_BYTES, numOfRows * pColumn->info.bytes); if (tmp == NULL) { return TSDB_CODE_OUT_OF_MEMORY; } + // copy back the existed data + if (pColumn->pData != NULL) { + memcpy(tmp, pColumn->pData, existedRows * pColumn->info.bytes); + taosMemoryFreeClear(pColumn->pData); + } + pColumn->pData = tmp; if (clearPayload) { memset(tmp + pColumn->info.bytes * existedRows, 0, pColumn->info.bytes * (numOfRows - existedRows)); diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index ed297e2b66..a98b172873 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -271,67 +271,6 @@ static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool is return v; } -//static int64_t i64VectorCmpAVX2(const int64_t* pData, int32_t numOfRows, bool isMinFunc) { -// int64_t v = 0; -// const int32_t bitWidth = 256; -// const int64_t* p = pData; -// -// int32_t width = (bitWidth>>3u) / sizeof(int64_t); -// int32_t remain = numOfRows % width; -// int32_t rounds = numOfRows / width; -// -//#if __AVX2__ -// __m256i next; -// __m256i initialVal = _mm256_loadu_si256((__m256i*)p); -// p += width; -// -// if (!isMinFunc) { // max function -// for (int32_t i = 0; i < rounds; ++i) { -// next = _mm256_lddqu_si256((__m256i*)p); -// initialVal = _mm256_max_epi64(initialVal, next); -// p += width; -// } -// -// // let sum up the final results -// const int64_t* q = (const int64_t*)&initialVal; -// v = TMAX(q[0], q[1]); -// for(int32_t k = 1; k < width; ++k) { -// v = TMAX(v, q[k]); -// } -// -// // calculate the front and the reminder items in array list -// int32_t start = rounds * width; -// for (int32_t j = 0; j < remain; ++j) { -// if (v < p[j + start]) { -// v = p[j + start]; -// } -// } -// } else { // min function -// for (int32_t i = 0; i < rounds; ++i) { -// next = _mm256_lddqu_si256((__m256i*)p); -// initialVal = _mm256_min_epi64(initialVal, next); -// p += width; -// } -// -// // let sum up the final results -// const int64_t* q = (const int64_t*)&initialVal; -// v = TMIN(q[0], q[1]); -// for(int32_t k = 1; k < width; ++k) { -// v = TMIN(v, q[k]); -// } -// -// // calculate the front and the remainder items in array list -// int32_t start = rounds * width; -// for (int32_t j = 0; j < remain; ++j) { -// if (v > p[j + start]) { -// v = p[j + start]; -// } -// } -// } -//#endif -// -// return v; -//} static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { diff --git a/source/os/src/osMemory.c b/source/os/src/osMemory.c index 78fa362179..1facff1f3b 100644 --- a/source/os/src/osMemory.c +++ b/source/os/src/osMemory.c @@ -345,3 +345,11 @@ void taosMemoryTrim(int32_t size) { malloc_trim(size); #endif } + +void* taosMemoryMallocAlign(uint32_t alignment, int64_t size) { +#ifdef USE_TD_MEMORY + ASSERT(0); +#else + return memalign(alignment, size); +#endif +} From e0e55b2d6510e8d1420a8bc8852a9a9f0797c495 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Wed, 23 Nov 2022 22:08:35 +0800 Subject: [PATCH 38/83] refactor: do some internal refactor. --- include/common/tdatablock.h | 13 -------- include/common/tmisce.h | 42 +++++++++++++++++++++++++ source/client/inc/clientInt.h | 2 +- source/client/src/clientMain.c | 2 +- source/client/src/clientMsgHandler.c | 1 + source/common/src/tglobal.c | 2 +- source/common/src/tmisce.c | 2 +- source/dnode/mgmt/node_util/src/dmEps.c | 1 + source/dnode/mnode/impl/src/mndDnode.c | 1 + source/dnode/mnode/impl/src/mndMnode.c | 2 +- source/dnode/mnode/impl/src/mndVgroup.c | 1 + 11 files changed, 51 insertions(+), 18 deletions(-) create mode 100644 include/common/tmisce.h diff --git a/include/common/tdatablock.h b/include/common/tdatablock.h index 502ba10d33..76ab2f7c3f 100644 --- a/include/common/tdatablock.h +++ b/include/common/tdatablock.h @@ -24,11 +24,6 @@ extern "C" { #endif -typedef struct SCorEpSet { - int32_t version; - SEpSet epSet; -} SCorEpSet; - typedef struct SBlockOrderInfo { bool nullFirst; int32_t order; @@ -36,14 +31,6 @@ typedef struct SBlockOrderInfo { SColumnInfoData* pColData; } SBlockOrderInfo; -int32_t taosGetFqdnPortFromEp(const char* ep, SEp* pEp); -void addEpIntoEpSet(SEpSet* pEpSet, const char* fqdn, uint16_t port); - -bool isEpsetEqual(const SEpSet* s1, const SEpSet* s2); - -void updateEpSet_s(SCorEpSet* pEpSet, SEpSet* pNewEpSet); -SEpSet getEpSet_s(SCorEpSet* pEpSet); - #define NBIT (3u) #define BitPos(_n) ((_n) & ((1 << NBIT) - 1)) #define BMCharPos(bm_, r_) ((bm_)[(r_) >> NBIT]) diff --git a/include/common/tmisce.h b/include/common/tmisce.h new file mode 100644 index 0000000000..b9f5cf5b91 --- /dev/null +++ b/include/common/tmisce.h @@ -0,0 +1,42 @@ +/* + * Copyright (c) 2019 TAOS Data, Inc. + * + * This program is free software: you can use, redistribute, and/or modify + * it under the terms of the GNU Affero General Public License, version 3 + * or later ("AGPL"), as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. + * + * You should have received a copy of the GNU Affero General Public License + * along with this program. If not, see . + */ + +#ifndef TDENGINE_TMISCE_H +#define TDENGINE_TMISCE_H + +#ifdef __cplusplus +extern "C" { +#endif + +#include "tmsg.h" + +typedef struct SCorEpSet { + int32_t version; + SEpSet epSet; +} SCorEpSet; + +int32_t taosGetFqdnPortFromEp(const char* ep, SEp* pEp); +void addEpIntoEpSet(SEpSet* pEpSet, const char* fqdn, uint16_t port); + +bool isEpsetEqual(const SEpSet* s1, const SEpSet* s2); + +void updateEpSet_s(SCorEpSet* pEpSet, SEpSet* pNewEpSet); +SEpSet getEpSet_s(SCorEpSet* pEpSet); + +#ifdef __cplusplus +} +#endif + +#endif // TDENGINE_TMISCE_H diff --git a/source/client/inc/clientInt.h b/source/client/inc/clientInt.h index e569a97723..aae20c587d 100644 --- a/source/client/inc/clientInt.h +++ b/source/client/inc/clientInt.h @@ -26,7 +26,7 @@ extern "C" { #include "query.h" #include "taos.h" #include "tcommon.h" -#include "tdatablock.h" +#include "tmisce.h" #include "tdef.h" #include "thash.h" #include "tlist.h" diff --git a/source/client/src/clientMain.c b/source/client/src/clientMain.c index efa7d095c5..a84f9e6bb0 100644 --- a/source/client/src/clientMain.c +++ b/source/client/src/clientMain.c @@ -20,13 +20,13 @@ #include "functionMgt.h" #include "os.h" #include "query.h" -#include "qworker.h" #include "scheduler.h" #include "tglobal.h" #include "tmsg.h" #include "tref.h" #include "trpc.h" #include "version.h" +#include "tdatablock.h" #define TSC_VAR_NOT_RELEASE 1 #define TSC_VAR_RELEASED 0 diff --git a/source/client/src/clientMsgHandler.c b/source/client/src/clientMsgHandler.c index 591d469d7c..a49d072569 100644 --- a/source/client/src/clientMsgHandler.c +++ b/source/client/src/clientMsgHandler.c @@ -20,6 +20,7 @@ #include "query.h" #include "tdef.h" #include "tname.h" +#include "tdatablock.h" static void setErrno(SRequestObj* pRequest, int32_t code) { pRequest->code = code; diff --git a/source/common/src/tglobal.c b/source/common/src/tglobal.c index 0db3e99863..d82ad5c15a 100644 --- a/source/common/src/tglobal.c +++ b/source/common/src/tglobal.c @@ -16,7 +16,7 @@ #define _DEFAULT_SOURCE #include "tglobal.h" #include "tconfig.h" -#include "tdatablock.h" +#include "tmisce.h" #include "tgrant.h" #include "tlog.h" diff --git a/source/common/src/tmisce.c b/source/common/src/tmisce.c index dfaebc99f6..dfb1917fcf 100644 --- a/source/common/src/tmisce.c +++ b/source/common/src/tmisce.c @@ -14,7 +14,7 @@ */ #define _DEFAULT_SOURCE -#include "tdatablock.h" +#include "tmisce.h" #include "tglobal.h" #include "tlog.h" #include "tname.h" diff --git a/source/dnode/mgmt/node_util/src/dmEps.c b/source/dnode/mgmt/node_util/src/dmEps.c index 7fe7d44827..4bc2e64fad 100644 --- a/source/dnode/mgmt/node_util/src/dmEps.c +++ b/source/dnode/mgmt/node_util/src/dmEps.c @@ -15,6 +15,7 @@ #define _DEFAULT_SOURCE #include "dmUtil.h" +#include "tmisce.h" static void dmPrintEps(SDnodeData *pData); static bool dmIsEpChanged(SDnodeData *pData, int32_t dnodeId, const char *ep); diff --git a/source/dnode/mnode/impl/src/mndDnode.c b/source/dnode/mnode/impl/src/mndDnode.c index 521f924fad..2a3ecf1924 100644 --- a/source/dnode/mnode/impl/src/mndDnode.c +++ b/source/dnode/mnode/impl/src/mndDnode.c @@ -24,6 +24,7 @@ #include "mndTrans.h" #include "mndUser.h" #include "mndVgroup.h" +#include "tmisce.h" #define TSDB_DNODE_VER_NUMBER 1 #define TSDB_DNODE_RESERVE_SIZE 64 diff --git a/source/dnode/mnode/impl/src/mndMnode.c b/source/dnode/mnode/impl/src/mndMnode.c index 563b3cd3cf..b02ca4c773 100644 --- a/source/dnode/mnode/impl/src/mndMnode.c +++ b/source/dnode/mnode/impl/src/mndMnode.c @@ -20,7 +20,7 @@ #include "mndShow.h" #include "mndSync.h" #include "mndTrans.h" -#include "mndUser.h" +#include "tmisce.h" #define MNODE_VER_NUMBER 1 #define MNODE_RESERVE_SIZE 64 diff --git a/source/dnode/mnode/impl/src/mndVgroup.c b/source/dnode/mnode/impl/src/mndVgroup.c index e00d0d955e..5555335765 100644 --- a/source/dnode/mnode/impl/src/mndVgroup.c +++ b/source/dnode/mnode/impl/src/mndVgroup.c @@ -22,6 +22,7 @@ #include "mndShow.h" #include "mndTrans.h" #include "mndUser.h" +#include "tmisce.h" #define VGROUP_VER_NUMBER 1 #define VGROUP_RESERVE_SIZE 64 From d71439260fc7353ef0aea8689a5e3dc995ce83aa Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 24 Nov 2022 15:01:25 +0800 Subject: [PATCH 39/83] refactor(query): add simd support for minmax query. --- .../libs/function/src/detail/tavgfunction.c | 33 -- source/libs/function/src/detail/tminmax.c | 282 +++++++++++++++++- 2 files changed, 268 insertions(+), 47 deletions(-) diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 744927d6c8..1553a446a7 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -272,39 +272,6 @@ static void i64VectorSumAVX2(const int64_t* plist, int32_t numOfRows, SAvgRes* p #endif } -static int32_t handleFloatCols(const SColumnInfoData* pCol, const SInputColumnInfoData* pInput, SAvgRes* pRes) { - int32_t numOfElems = 0; - float* plist = (float*)pCol->pData; - - const int32_t THRESHOLD_SIZE = 8; - - if (pCol->hasNull || pInput->numOfRows <= THRESHOLD_SIZE) { - for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - numOfElems += 1; - pRes->count += 1; - pRes->sum.dsum += plist[i]; - } - } else { // no null values exist - numOfElems = pInput->numOfRows; - pRes->count += pInput->numOfRows; - - // 3. If the CPU supports AVX, let's employ AVX instructions to speedup this loop - if (tsAVXEnable && tsSIMDEnable) { - floatVectorSumAVX(plist, pInput->numOfRows, pRes); - } else { - for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) { - pRes->sum.dsum += plist[i]; - } - } - } - - return numOfElems; -} - int32_t getAvgInfoSize() { return (int32_t)sizeof(SAvgRes); } bool getAvgFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index a98b172873..bda1fea90a 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -30,7 +30,7 @@ static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool is #if __AVX2__ __m256i next; - __m256i initialVal = _mm256_loadu_si256((__m256i*)p); + __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); p += width; if (!isMinFunc) { // max function @@ -40,7 +40,7 @@ static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool is p += width; } - // let sum up the final results + // let compare the final results const int32_t* q = (const int32_t*)&initialVal; v = TMAX(q[0], q[1]); for (int32_t k = 1; k < width; ++k) { @@ -155,7 +155,7 @@ static int8_t i8VectorCmpAVX2(const int8_t* pData, int32_t numOfRows, bool isMin #if __AVX2__ __m256i next; - __m256i initialVal = _mm256_loadu_si256((__m256i*)p); + __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); p += width; if (!isMinFunc) { // max function @@ -218,7 +218,7 @@ static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool is #if __AVX2__ __m256i next; - __m256i initialVal = _mm256_loadu_si256((__m256i*)p); + __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); p += width; if (!isMinFunc) { // max function @@ -271,6 +271,179 @@ static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool is return v; } +static int32_t handleInt8Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, + SMinmaxResInfo* pBuf, bool isMinFunc) { + int8_t* pData = (int8_t*)pCol->pData; + int8_t* val = (int8_t*)&pBuf->v; + + int32_t numOfElems = 0; + if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { + int32_t i = start; + while (i < (start + numOfRows)) { + if (!colDataIsNull_f(pCol->nullbitmap, i)) { + break; + } + i += 1; + } + + if ((i < (start + numOfRows)) && (!pBuf->assign)) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + numOfElems += 1; + } + + if (isMinFunc) { // min + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } else { // max function + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } + } else { // not has null value + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + *val = i8VectorCmpAVX2(pData, numOfRows, isMinFunc); + } else { + if (!pBuf->assign) { + *val = pData[0]; + pBuf->assign = true; + } + + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val > pData[i]) { + *val = pData[i]; + } + } + } else { // max + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val < pData[i]) { + *val = pData[i]; + } + } + } + } + + numOfElems = numOfRows; + } + + return numOfElems; +} + +static int32_t handleInt16Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, + SMinmaxResInfo* pBuf, bool isMinFunc) { + int16_t* pData = (int16_t*)pCol->pData; + int16_t* val = (int16_t*)&pBuf->v; + + int32_t numOfElems = 0; + if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { + int32_t i = start; + while (i < (start + numOfRows)) { + if (!colDataIsNull_f(pCol->nullbitmap, i)) { + break; + } + i += 1; + } + + if ((i < (start + numOfRows)) && (!pBuf->assign)) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + numOfElems += 1; + } + + if (isMinFunc) { // min + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } else { // max function + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } + } else { // not has null value + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + *val = i16VectorCmpAVX2(pData, numOfRows, isMinFunc); + } else { + if (!pBuf->assign) { + *val = pData[0]; + pBuf->assign = true; + } + + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val > pData[i]) { + *val = pData[i]; + } + } + } else { // max + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val < pData[i]) { + *val = pData[i]; + } + } + } + } + + numOfElems = numOfRows; + } + + return numOfElems; +} static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { @@ -359,6 +532,87 @@ static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numO return numOfElems; } +static int32_t handleInt64Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, + SMinmaxResInfo* pBuf, bool isMinFunc) { + int32_t* pData = (int32_t*)pCol->pData; + int32_t* val = (int32_t*)&pBuf->v; + + int32_t numOfElems = 0; + if (pCol->hasNull || pCtx->subsidiaries.num > 0) { + int32_t i = start; + while (i < (start + numOfRows)) { + if (!colDataIsNull_f(pCol->nullbitmap, i)) { + break; + } + i += 1; + } + + if ((i < (start + numOfRows)) && (!pBuf->assign)) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + numOfElems += 1; + } + + if (isMinFunc) { // min + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + + if (*val > pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + + } else { // max function + for (; i < start + numOfRows; ++i) { + if (colDataIsNull_f(pCol->nullbitmap, i)) { + continue; + } + // ignore the equivalent data value + // NOTE: An faster version to avoid one additional comparison with FPU. + if (*val < pData[i]) { + *val = pData[i]; + if (pCtx->subsidiaries.num > 0) { + updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); + } + } + numOfElems += 1; + } + } + } else { // not has null value + // AVX2 version to speedup the loop + if (!pBuf->assign) { + *val = pData[0]; + pBuf->assign = true; + } + + if (isMinFunc) { // min + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val > pData[i]) { + *val = pData[i]; + } + } + } else { // max + for (int32_t i = start; i < start + numOfRows; ++i) { + if (*val < pData[i]) { + *val = pData[i]; + } + } + } + + numOfElems = numOfRows; + } + return numOfElems; +} + static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { float* pData = (float*)pCol->pData; @@ -445,13 +699,13 @@ static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numO return numOfElems; } -static int32_t handleInt8Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, - SMinmaxResInfo* pBuf, bool isMinFunc) { - int8_t* pData = (int8_t*)pCol->pData; - int8_t* val = (int8_t*)&pBuf->v; +static int32_t handleDoubleCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, + SMinmaxResInfo* pBuf, bool isMinFunc) { + float* pData = (float*)pCol->pData; + double* val = (double*)&pBuf->v; int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { + if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { int32_t i = start; while (i < (start + numOfRows)) { if (!colDataIsNull_f(pCol->nullbitmap, i)) { @@ -483,12 +737,12 @@ static int32_t handleInt8Col(SColumnInfoData* pCol, int32_t start, int32_t numOf } numOfElems += 1; } - } else { // max function for (; i < start + numOfRows; ++i) { if (colDataIsNull_f(pCol->nullbitmap, i)) { continue; } + // ignore the equivalent data value // NOTE: An faster version to avoid one additional comparison with FPU. if (*val < pData[i]) { @@ -499,12 +753,11 @@ static int32_t handleInt8Col(SColumnInfoData* pCol, int32_t start, int32_t numOf } numOfElems += 1; } - } } else { // not has null value - // AVX2 version to speedup the loop - if (tsAVX2Enable && tsSIMDEnable) { - *val = i8VectorCmpAVX2(pData, numOfRows, isMinFunc); + // AVX version to speedup the loop + if (tsAVXEnable && tsSIMDEnable) { + *val = (double) floatVectorCmpAVX(pData, numOfRows, isMinFunc); } else { if (!pBuf->assign) { *val = pData[0]; @@ -660,6 +913,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) { numOfElems = handleInt8Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); } else if (type == TSDB_DATA_TYPE_SMALLINT) { + numOfElems = handleInt16Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); int16_t* pData = (int16_t*)pCol->pData; int16_t* val = (int16_t*)&pBuf->v; From 0ed2c923ad54efebbe2cc9a9286c300b745986a9 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 24 Nov 2022 17:01:00 +0800 Subject: [PATCH 40/83] enh(query): support simd in min/max query. --- source/libs/function/src/detail/tminmax.c | 551 +++++++++------------- source/libs/function/src/udfd.c | 1 + 2 files changed, 218 insertions(+), 334 deletions(-) diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index bda1fea90a..21f8631156 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -19,139 +19,20 @@ #include "tfunctionInt.h" #include "tglobal.h" -static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool isMinFunc) { - int32_t v = 0; - const int32_t bitWidth = 256; - const int32_t* p = pData; - - int32_t width = (bitWidth>>3u) / sizeof(int32_t); - int32_t remain = numOfRows % width; - int32_t rounds = numOfRows / width; - -#if __AVX2__ - __m256i next; - __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); - p += width; - - if (!isMinFunc) { // max function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_max_epi32(initialVal, next); - p += width; - } - - // let compare the final results - const int32_t* q = (const int32_t*)&initialVal; - v = TMAX(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMAX(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + start]) { - v = p[j + start]; - } - } - } else { // min function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_min_epi32(initialVal, next); - p += width; - } - - // let sum up the final results - const int32_t* q = (const int32_t*)&initialVal; - v = TMIN(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMIN(v, q[k]); - } - - // calculate the front and the remainder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + start]) { - v = p[j + start]; - } - } - } -#endif - - return v; -} - -static float floatVectorCmpAVX(const float* pData, int32_t numOfRows, bool isMinFunc) { - float v = 0; +static void calculateRounds(int32_t numOfRows, int32_t bytes, int32_t* remainder, int32_t* rounds, int32_t* width) { const int32_t bitWidth = 256; - const float* p = pData; - int32_t width = (bitWidth>>3u) / sizeof(float); - int32_t remain = numOfRows % width; - int32_t rounds = numOfRows / width; - -#if __AVX__ - - __m256 next; - __m256 initialVal = _mm256_loadu_ps(p); - p += width; - - if (!isMinFunc) { // max function - for (int32_t i = 1; i < rounds; ++i) { - next = _mm256_loadu_ps(p); - initialVal = _mm256_max_ps(initialVal, next); - p += width; - } - - // let sum up the final results - const float* q = (const float*)&initialVal; - v = TMAX(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMAX(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + width]) { - v = p[j + width]; - } - } - } else { // min function - for (int32_t i = 1; i < rounds; ++i) { - next = _mm256_loadu_ps(p); - initialVal = _mm256_min_ps(initialVal, next); - p += width; - } - - // let sum up the final results - const float* q = (const float*)&initialVal; - v = TMIN(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMIN(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * bitWidth; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + start]) { - v = p[j + start]; - } - } - } -#endif - - return v; + *width = (bitWidth>>3u) / bytes; + *remainder = numOfRows % (*width); + *rounds = numOfRows / (*width); } static int8_t i8VectorCmpAVX2(const int8_t* pData, int32_t numOfRows, bool isMinFunc) { int8_t v = 0; - const int32_t bitWidth = 256; const int8_t* p = pData; - int32_t width = (bitWidth>>3u) / sizeof(int8_t); - int32_t remain = numOfRows % width; - int32_t rounds = numOfRows / width; + int32_t width, remain, rounds; + calculateRounds(numOfRows, sizeof(int8_t), &remain, &rounds, &width); #if __AVX2__ __m256i next; @@ -209,12 +90,10 @@ static int8_t i8VectorCmpAVX2(const int8_t* pData, int32_t numOfRows, bool isMin static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool isMinFunc) { int16_t v = 0; - const int32_t bitWidth = 256; const int16_t* p = pData; - int32_t width = (bitWidth>>3u) / sizeof(int16_t); - int32_t remain = numOfRows % width; - int32_t rounds = numOfRows / width; + int32_t width, remain, rounds; + calculateRounds(numOfRows, sizeof(int16_t), &remain, &rounds, &width); #if __AVX2__ __m256i next; @@ -236,6 +115,7 @@ static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool is v = TMAX(v, q[k]); } + // calculate the front and the reminder items in array list int32_t start = rounds * width; for (int32_t j = 0; j < remain; ++j) { @@ -271,20 +151,208 @@ static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool is return v; } +static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool isMinFunc) { + int32_t v = 0; + const int32_t* p = pData; + + int32_t width, remain, rounds; + calculateRounds(numOfRows, sizeof(int32_t), &remain, &rounds, &width); + +#if __AVX2__ + __m256i next; + __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); + p += width; + + if (!isMinFunc) { // max function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initialVal = _mm256_max_epi32(initialVal, next); + p += width; + } + + // let compare the final results + const int32_t* q = (const int32_t*)&initialVal; + v = TMAX(q[0], q[1]); + for (int32_t k = 1; k < width; ++k) { + v = TMAX(v, q[k]); + } + + // calculate the front and the reminder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + start]) { + v = p[j + start]; + } + } + } else { // min function + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initialVal = _mm256_min_epi32(initialVal, next); + p += width; + } + + // let sum up the final results + const int32_t* q = (const int32_t*)&initialVal; + v = TMIN(q[0], q[1]); + for (int32_t k = 1; k < width; ++k) { + v = TMIN(v, q[k]); + } + + // calculate the front and the remainder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + start]) { + v = p[j + start]; + } + } + } +#endif + + return v; +} + +static float floatVectorCmpAVX(const float* pData, int32_t numOfRows, bool isMinFunc) { + float v = 0; + const float* p = pData; + + int32_t width, remain, rounds; + calculateRounds(numOfRows, sizeof(float), &remain, &rounds, &width); + +#if __AVX__ + + __m256 next; + __m256 initialVal = _mm256_loadu_ps(p); + p += width; + + if (!isMinFunc) { // max function + for (int32_t i = 1; i < rounds; ++i) { + next = _mm256_loadu_ps(p); + initialVal = _mm256_max_ps(initialVal, next); + p += width; + } + + // let sum up the final results + const float* q = (const float*)&initialVal; + v = TMAX(q[0], q[1]); + for (int32_t k = 1; k < width; ++k) { + v = TMAX(v, q[k]); + } + + // calculate the front and the reminder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + start]) { + v = p[j + start]; + } + } + } else { // min function + for (int32_t i = 1; i < rounds; ++i) { + next = _mm256_loadu_ps(p); + initialVal = _mm256_min_ps(initialVal, next); + p += width; + } + + // let sum up the final results + const float* q = (const float*)&initialVal; + v = TMIN(q[0], q[1]); + for (int32_t k = 1; k < width; ++k) { + v = TMIN(v, q[k]); + } + + // calculate the front and the reminder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + start]) { + v = p[j + start]; + } + } + } +#endif + + return v; +} + +static double doubleVectorCmpAVX(const double* pData, int32_t numOfRows, bool isMinFunc) { + double v = 0; + const double* p = pData; + + int32_t width, remain, rounds; + calculateRounds(numOfRows, sizeof(double), &remain, &rounds, &width); + +#if __AVX__ + + __m256d next; + __m256d initialVal = _mm256_loadu_pd(p); + p += width; + + if (!isMinFunc) { // max function + for (int32_t i = 1; i < rounds; ++i) { + next = _mm256_loadu_pd(p); + initialVal = _mm256_max_pd(initialVal, next); + p += width; + } + + // let sum up the final results + const double* q = (const double*)&initialVal; + v = TMAX(q[0], q[1]); + for (int32_t k = 1; k < width; ++k) { + v = TMAX(v, q[k]); + } + + // calculate the front and the reminder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v < p[j + start]) { + v = p[j + start]; + } + } + } else { // min function + for (int32_t i = 1; i < rounds; ++i) { + next = _mm256_loadu_pd(p); + initialVal = _mm256_min_pd(initialVal, next); + p += width; + } + + // let sum up the final results + const double* q = (const double*)&initialVal; + v = TMIN(q[0], q[1]); + for (int32_t k = 1; k < width; ++k) { + v = TMIN(v, q[k]); + } + + // calculate the front and the reminder items in array list + int32_t start = rounds * width; + for (int32_t j = 0; j < remain; ++j) { + if (v > p[j + start]) { + v = p[j + start]; + } + } + } +#endif + + return v; +} + +static int32_t findFirstVal(const SColumnInfoData* pCol, int32_t start, int32_t numOfRows) { + int32_t i = start; + while (i < (start + numOfRows)) { + if (!colDataIsNull_f(pCol->nullbitmap, i)) { + break; + } + i += 1; + } + + return i; +} + static int32_t handleInt8Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { int8_t* pData = (int8_t*)pCol->pData; int8_t* val = (int8_t*)&pBuf->v; int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { - int32_t i = start; - while (i < (start + numOfRows)) { - if (!colDataIsNull_f(pCol->nullbitmap, i)) { - break; - } - i += 1; - } + if (pCol->hasNull || numOfRows <= 32 || pCtx->subsidiaries.num > 0) { + int32_t i = findFirstVal(pCol, start, numOfRows); if ((i < (start + numOfRows)) && (!pBuf->assign)) { *val = pData[i]; @@ -365,13 +433,7 @@ static int32_t handleInt16Col(SColumnInfoData* pCol, int32_t start, int32_t numO int32_t numOfElems = 0; if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { - int32_t i = start; - while (i < (start + numOfRows)) { - if (!colDataIsNull_f(pCol->nullbitmap, i)) { - break; - } - i += 1; - } + int32_t i = findFirstVal(pCol, start, numOfRows); if ((i < (start + numOfRows)) && (!pBuf->assign)) { *val = pData[i]; @@ -452,13 +514,7 @@ static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numO int32_t numOfElems = 0; if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { - int32_t i = start; - while (i < (start + numOfRows)) { - if (!colDataIsNull_f(pCol->nullbitmap, i)) { - break; - } - i += 1; - } + int32_t i = findFirstVal(pCol, start, numOfRows); if ((i < (start + numOfRows)) && (!pBuf->assign)) { *val = pData[i]; @@ -539,13 +595,7 @@ static int32_t handleInt64Col(SColumnInfoData* pCol, int32_t start, int32_t numO int32_t numOfElems = 0; if (pCol->hasNull || pCtx->subsidiaries.num > 0) { - int32_t i = start; - while (i < (start + numOfRows)) { - if (!colDataIsNull_f(pCol->nullbitmap, i)) { - break; - } - i += 1; - } + int32_t i = findFirstVal(pCol, start, numOfRows); if ((i < (start + numOfRows)) && (!pBuf->assign)) { *val = pData[i]; @@ -616,18 +666,11 @@ static int32_t handleInt64Col(SColumnInfoData* pCol, int32_t start, int32_t numO static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { float* pData = (float*)pCol->pData; - double* val = (double*)&pBuf->v; + float* val = (float*)&pBuf->v; int32_t numOfElems = 0; if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { - int32_t i = start; - while (i < (start + numOfRows)) { - if (!colDataIsNull_f(pCol->nullbitmap, i)) { - break; - } - i += 1; - } - + int32_t i = findFirstVal(pCol, start, numOfRows); if ((i < (start + numOfRows)) && (!pBuf->assign)) { *val = pData[i]; if (pCtx->subsidiaries.num > 0) { @@ -701,18 +744,12 @@ static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numO static int32_t handleDoubleCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, bool isMinFunc) { - float* pData = (float*)pCol->pData; + double* pData = (double*)pCol->pData; double* val = (double*)&pBuf->v; int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { - int32_t i = start; - while (i < (start + numOfRows)) { - if (!colDataIsNull_f(pCol->nullbitmap, i)) { - break; - } - i += 1; - } + if (pCol->hasNull || numOfRows < 4 || pCtx->subsidiaries.num > 0) { + int32_t i = findFirstVal(pCol, start, numOfRows); if ((i < (start + numOfRows)) && (!pBuf->assign)) { *val = pData[i]; @@ -757,7 +794,7 @@ static int32_t handleDoubleCol(SColumnInfoData* pCol, int32_t start, int32_t num } else { // not has null value // AVX version to speedup the loop if (tsAVXEnable && tsSIMDEnable) { - *val = (double) floatVectorCmpAVX(pData, numOfRows, isMinFunc); + *val = (double) doubleVectorCmpAVX(pData, numOfRows, isMinFunc); } else { if (!pBuf->assign) { *val = pData[0]; @@ -813,7 +850,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { if (IS_NULL_TYPE(type)) { numOfElems = 0; - goto _min_max_over; + goto _over; } // data in current data block are qualified to the query @@ -914,117 +951,10 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { numOfElems = handleInt8Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); } else if (type == TSDB_DATA_TYPE_SMALLINT) { numOfElems = handleInt16Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); - int16_t* pData = (int16_t*)pCol->pData; - int16_t* val = (int16_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } } else if (type == TSDB_DATA_TYPE_INT) { numOfElems = handleInt32Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); -#if 0 - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } -#endif - } else if (type == TSDB_DATA_TYPE_BIGINT) { - int64_t* pData = (int64_t*)pCol->pData; - int64_t* val = (int64_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } + numOfElems = handleInt64Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); } } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { if (type == TSDB_DATA_TYPE_UTINYINT) { @@ -1215,56 +1145,9 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { } } else if (type == TSDB_DATA_TYPE_FLOAT) { numOfElems = handleFloatCol(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); -#if 0 - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { -#if 0 - if ((*val) == pData[i]) { - continue; - } - - if ((*val < pData[i]) ^ isMinFunc) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } -#endif - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; - } -#endif - } -_min_max_over: +_over: if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); pBuf->nullTupleSaved = true; diff --git a/source/libs/function/src/udfd.c b/source/libs/function/src/udfd.c index 088aa62248..2ad36469ff 100644 --- a/source/libs/function/src/udfd.c +++ b/source/libs/function/src/udfd.c @@ -27,6 +27,7 @@ #include "tglobal.h" #include "tmsg.h" #include "trpc.h" +#include "tmisce.h" // clang-foramt on typedef struct SUdfdContext { From 89341bc303375abdfb0adacc409cd300a9125971 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 24 Nov 2022 17:01:37 +0800 Subject: [PATCH 41/83] enh(query): support simd in min/max query. --- source/libs/function/src/detail/tminmax.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index 21f8631156..5fbc70a213 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -432,7 +432,7 @@ static int32_t handleInt16Col(SColumnInfoData* pCol, int32_t start, int32_t numO int16_t* val = (int16_t*)&pBuf->v; int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { + if (pCol->hasNull || numOfRows <= 16 || pCtx->subsidiaries.num > 0) { int32_t i = findFirstVal(pCol, start, numOfRows); if ((i < (start + numOfRows)) && (!pBuf->assign)) { From f745dec5ec8f45dbeb805da1404e4491a7731b95 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 24 Nov 2022 17:17:22 +0800 Subject: [PATCH 42/83] other: merge 3.0 --- source/libs/executor/inc/executorimpl.h | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/source/libs/executor/inc/executorimpl.h b/source/libs/executor/inc/executorimpl.h index 10a322a758..fad232a7c0 100644 --- a/source/libs/executor/inc/executorimpl.h +++ b/source/libs/executor/inc/executorimpl.h @@ -153,6 +153,11 @@ typedef struct { SSchemaWrapper* qsw; } SSchemaInfo; +typedef struct { + SRWLatch lock; + SArray* pStopInfo; +} STaskStopInfo; + typedef struct SExecTaskInfo { STaskIdInfo id; uint32_t status; From 8dd780b486c45a2209a33a02581d51b3b7faa448 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 24 Nov 2022 17:17:53 +0800 Subject: [PATCH 43/83] other: merge 3.0 --- source/libs/executor/inc/executorimpl.h | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/source/libs/executor/inc/executorimpl.h b/source/libs/executor/inc/executorimpl.h index fad232a7c0..af3399bc2c 100644 --- a/source/libs/executor/inc/executorimpl.h +++ b/source/libs/executor/inc/executorimpl.h @@ -153,6 +153,11 @@ typedef struct { SSchemaWrapper* qsw; } SSchemaInfo; +typedef struct { + int32_t operatorType; + int64_t refId; +} SExchangeOpStopInfo; + typedef struct { SRWLatch lock; SArray* pStopInfo; From e5e817ae5539f8901180d460db77f7d06ea38832 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 24 Nov 2022 17:24:24 +0800 Subject: [PATCH 44/83] other: merge 3.0 --- source/libs/executor/inc/executorimpl.h | 8 ++++---- source/libs/executor/src/exchangeoperator.c | 10 +++++----- source/libs/function/inc/builtinsimpl.h | 2 ++ 3 files changed, 11 insertions(+), 9 deletions(-) diff --git a/source/libs/executor/inc/executorimpl.h b/source/libs/executor/inc/executorimpl.h index af3399bc2c..b9d622a4c3 100644 --- a/source/libs/executor/inc/executorimpl.h +++ b/source/libs/executor/inc/executorimpl.h @@ -153,17 +153,17 @@ typedef struct { SSchemaWrapper* qsw; } SSchemaInfo; -typedef struct { +typedef struct SExchangeOpStopInfo { int32_t operatorType; int64_t refId; } SExchangeOpStopInfo; -typedef struct { +typedef struct STaskStopInfo { SRWLatch lock; SArray* pStopInfo; } STaskStopInfo; -typedef struct SExecTaskInfo { +struct SExecTaskInfo { STaskIdInfo id; uint32_t status; STimeWindow window; @@ -260,7 +260,7 @@ typedef struct SExchangeInfo { // SArray, result block list, used to keep the multi-block that // passed by downstream operator - SArray* pReadyBlocks; + SArray* pResultBlockList; SArray* pRecycledBlocks;// build a pool for small data block to avoid to repeatly create and then destroy. SSDataBlock* pDummyBlock; // dummy block, not keep data bool seqLoadData; // sequential load data or not, false by default diff --git a/source/libs/executor/src/exchangeoperator.c b/source/libs/executor/src/exchangeoperator.c index cf1f5aa290..52aa3db0fd 100644 --- a/source/libs/executor/src/exchangeoperator.c +++ b/source/libs/executor/src/exchangeoperator.c @@ -182,7 +182,7 @@ static SSDataBlock* doLoadRemoteDataImpl(SOperatorInfo* pOperator) { } // we have buffered retrieved datablock, return it directly - SSDataBlock** p = taosArrayPop(pExchangeInfo->pReadyBlocks); + SSDataBlock** p = taosArrayPop(pExchangeInfo->pResultBlockList); if (p != NULL) { taosArrayPush(pExchangeInfo->pRecycledBlocks, p); return *p; @@ -193,10 +193,10 @@ static SSDataBlock* doLoadRemoteDataImpl(SOperatorInfo* pOperator) { concurrentlyLoadRemoteDataImpl(pOperator, pExchangeInfo, pTaskInfo); } - if (taosArrayGetSize(pExchangeInfo->pReadyBlocks) == 0) { + if (taosArrayGetSize(pExchangeInfo->pResultBlockList) == 0) { return NULL; } else { - p = taosArrayPop(pExchangeInfo->pReadyBlocks); + p = taosArrayPop(pExchangeInfo->pResultBlockList); taosArrayPush(pExchangeInfo->pRecycledBlocks, p); return *p; } @@ -298,7 +298,7 @@ SOperatorInfo* createExchangeOperatorInfo(void* pTransporter, SExchangePhysiNode tsem_init(&pInfo->ready, 0, 0); pInfo->pDummyBlock = createResDataBlock(pExNode->node.pOutputDataBlockDesc); - pInfo->pReadyBlocks = taosArrayInit(64, POINTER_BYTES); + pInfo->pResultBlockList = taosArrayInit(64, POINTER_BYTES); pInfo->pRecycledBlocks = taosArrayInit(64, POINTER_BYTES); SExchangeOpStopInfo stopInfo = {QUERY_NODE_PHYSICAL_PLAN_EXCHANGE, pInfo->self}; @@ -346,7 +346,7 @@ void doDestroyExchangeOperatorInfo(void* param) { taosArrayDestroy(pExInfo->pSources); taosArrayDestroyEx(pExInfo->pSourceDataInfo, freeSourceDataInfo); - taosArrayDestroyEx(pExInfo->pReadyBlocks, freeBlock); + taosArrayDestroyEx(pExInfo->pResultBlockList, freeBlock); taosArrayDestroyEx(pExInfo->pRecycledBlocks, freeBlock); blockDataDestroy(pExInfo->pDummyBlock); diff --git a/source/libs/function/inc/builtinsimpl.h b/source/libs/function/inc/builtinsimpl.h index 2ec882d1de..307a82e256 100644 --- a/source/libs/function/inc/builtinsimpl.h +++ b/source/libs/function/inc/builtinsimpl.h @@ -30,6 +30,8 @@ typedef struct SSumRes { double dsum; }; int16_t type; + int64_t prevTs; + bool isPrevTsSet; } SSumRes; typedef struct SMinmaxResInfo { From f4f2774100eb39cba09f392e12dd05c3d7a19199 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Thu, 24 Nov 2022 23:23:05 +0800 Subject: [PATCH 45/83] refactor: do some internal refactor. --- .../libs/function/src/detail/tavgfunction.c | 1 + source/libs/function/src/detail/tminmax.c | 1300 +++++++---------- 2 files changed, 529 insertions(+), 772 deletions(-) diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 1553a446a7..9507380d38 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -551,6 +551,7 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { } break; } + case TSDB_DATA_TYPE_FLOAT: { const float* plist = (const float*) pCol->pData; diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index 5fbc70a213..03438350ed 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -19,6 +19,48 @@ #include "tfunctionInt.h" #include "tglobal.h" +#define __COMPARE_ACQUIRED_MAX(i, end, bm, _data, ctx, val, pos) \ + for (; i < (end); ++i) { \ + if (colDataIsNull_f(bm, i)) { \ + continue; \ + } \ + \ + if ((val) < (_data)[i]) { \ + (val) = (_data)[i]; \ + if ((ctx)->subsidiaries.num > 0) { \ + updateTupleData((ctx), i, (ctx)->pSrcBlock, pos); \ + } \ + } \ + } + +#define __COMPARE_ACQUIRED_MIN(i, end, bm, _data, ctx, val, pos) \ + for (; i < (end); ++i) { \ + if (colDataIsNull_f(bm, i)) { \ + continue; \ + } \ + \ + if ((val) > (_data)[i]) { \ + (val) = (_data)[i]; \ + if ((ctx)->subsidiaries.num > 0) { \ + updateTupleData((ctx), i, (ctx)->pSrcBlock, pos); \ + } \ + } \ + } + +#define __COMPARE_EXTRACT_MIN(start, end, val, _data) \ + for (int32_t i = (start); i < (end); ++i) { \ + if ((val) > (_data)[i]) { \ + (val) = (_data)[i]; \ + } \ + } + +#define __COMPARE_EXTRACT_MAX(start, end, val, _data) \ + for (int32_t i = (start); i < (end); ++i) { \ + if ((val) < (_data)[i]) { \ + (val) = (_data)[i]; \ + } \ + } + static void calculateRounds(int32_t numOfRows, int32_t bytes, int32_t* remainder, int32_t* rounds, int32_t* width) { const int32_t bitWidth = 256; @@ -27,7 +69,32 @@ static void calculateRounds(int32_t numOfRows, int32_t bytes, int32_t* remainder *rounds = numOfRows / (*width); } -static int8_t i8VectorCmpAVX2(const int8_t* pData, int32_t numOfRows, bool isMinFunc) { +#define EXTRACT_MAX_VAL(_first, _sec, _width, _remain, _v) \ + (_v) = TMAX((_first)[0], (_first)[1]); \ + for (int32_t k = 1; k < (_width); ++k) { \ + (_v) = TMAX((_v), (_first)[k]); \ + } \ + \ + for (int32_t j = 0; j < (_remain); ++j) { \ + if ((_v) < (_sec)[j]) { \ + (_v) = (_sec)[j]; \ + } \ + } + +#define EXTRACT_MIN_VAL(_first, _sec, _width, _remain, _v) \ + (_v) = TMIN((_first)[0], (_first)[1]); \ + for (int32_t k = 1; k < (_width); ++k) { \ + (_v) = TMIN((_v), (_first)[k]); \ + } \ + \ + for (int32_t j = 0; j < (_remain); ++j) { \ + if ((_v) > (_sec)[j]) { \ + (_v) = (_sec)[j]; \ + } \ + } + + +static int8_t i8VectorCmpAVX2(const void* pData, int32_t numOfRows, bool isMinFunc, bool signVal) { int8_t v = 0; const int8_t* p = pData; @@ -36,51 +103,51 @@ static int8_t i8VectorCmpAVX2(const int8_t* pData, int32_t numOfRows, bool isMin #if __AVX2__ __m256i next; - __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); + __m256i initVal = _mm256_lddqu_si256((__m256i*)p); p += width; if (!isMinFunc) { // max function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_max_epi8(initialVal, next); - p += width; - } - - // let sum up the final results - const int8_t* q = (const int8_t*)&initialVal; - v = TMAX(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMAX(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + start]) { - v = p[j + start]; + if (signVal) { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_max_epi8(initVal, next); + p += width; } + + const int8_t* q = (const int8_t*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) + } else { // unsigned value + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_max_epu8(initVal, next); + p += width; + } + + const uint8_t* q = (const uint8_t*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) } + } else { // min function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_min_epi8(initialVal, next); - p += width; - } - - // let sum up the final results - const int8_t* q = (const int8_t*)&initialVal; - - v = TMIN(q[0], q[1]); - for(int32_t k = 1; k < width; ++k) { - v = TMIN(v, q[k]); - } - - // calculate the front and the remainder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + start]) { - v = p[j + start]; + if (signVal) { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_min_epi8(initVal, next); + p += width; } + + // let sum up the final results + const int8_t* q = (const int8_t*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) + } else { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_min_epu8(initVal, next); + p += width; + } + + // let sum up the final results + const uint8_t* q = (const uint8_t*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) } } #endif @@ -88,7 +155,7 @@ static int8_t i8VectorCmpAVX2(const int8_t* pData, int32_t numOfRows, bool isMin return v; } -static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool isMinFunc) { +static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool isMinFunc, bool signVal) { int16_t v = 0; const int16_t* p = pData; @@ -97,53 +164,53 @@ static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool is #if __AVX2__ __m256i next; - __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); + __m256i initVal = _mm256_lddqu_si256((__m256i*)p); p += width; if (!isMinFunc) { // max function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_max_epi16(initialVal, next); - p += width; - } - - // let sum up the final results - const int16_t* q = (const int16_t*)&initialVal; - - v = TMAX(q[0], q[1]); - for(int32_t k = 1; k < width; ++k) { - v = TMAX(v, q[k]); - } - - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + start]) { - v = p[j + start]; + if (signVal) { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_max_epi16(initVal, next); + p += width; } + + // let sum up the final results + const int16_t* q = (const int16_t*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) + } else { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_max_epu16(initVal, next); + p += width; + } + + // let sum up the final results + const uint16_t* q = (const uint16_t*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) } + } else { // min function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_min_epi16(initialVal, next); - p += width; - } - - // let sum up the final results - const int16_t* q = (const int16_t*)&initialVal; - - v = TMIN(q[0], q[1]); - for(int32_t k = 1; k < width; ++k) { - v = TMIN(v, q[k]); - } - - // calculate the front and the remainder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + start]) { - v = p[j + start]; + if (signVal) { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_min_epi16(initVal, next); + p += width; } + + // let sum up the final results + const int16_t* q = (const int16_t*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) + } else { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_min_epi16(initVal, next); + p += width; + } + + // let sum up the final results + const uint16_t* q = (const uint16_t*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) } } #endif @@ -151,7 +218,7 @@ static int16_t i16VectorCmpAVX2(const int16_t* pData, int32_t numOfRows, bool is return v; } -static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool isMinFunc) { +static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool isMinFunc, bool signVal) { int32_t v = 0; const int32_t* p = pData; @@ -160,50 +227,52 @@ static int32_t i32VectorCmpAVX2(const int32_t* pData, int32_t numOfRows, bool is #if __AVX2__ __m256i next; - __m256i initialVal = _mm256_lddqu_si256((__m256i*)p); + __m256i initVal = _mm256_lddqu_si256((__m256i*)p); p += width; if (!isMinFunc) { // max function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_max_epi32(initialVal, next); - p += width; - } - - // let compare the final results - const int32_t* q = (const int32_t*)&initialVal; - v = TMAX(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMAX(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + start]) { - v = p[j + start]; + if (signVal) { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_max_epi32(initVal, next); + p += width; } + + // let compare the final results + const int32_t* q = (const int32_t*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) + } else { // unsigned value + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_max_epi32(initVal, next); + p += width; + } + + // let compare the final results + const uint32_t* q = (const uint32_t*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) } } else { // min function - for (int32_t i = 0; i < rounds; ++i) { - next = _mm256_lddqu_si256((__m256i*)p); - initialVal = _mm256_min_epi32(initialVal, next); - p += width; - } - - // let sum up the final results - const int32_t* q = (const int32_t*)&initialVal; - v = TMIN(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMIN(v, q[k]); - } - - // calculate the front and the remainder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + start]) { - v = p[j + start]; + if (signVal) { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_min_epi32(initVal, next); + p += width; } + + // let sum up the final results + const int32_t* q = (const int32_t*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) + } else { + for (int32_t i = 0; i < rounds; ++i) { + next = _mm256_lddqu_si256((__m256i*)p); + initVal = _mm256_min_epu32(initVal, next); + p += width; + } + + // let sum up the final results + const uint32_t* q = (const uint32_t*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) } } #endif @@ -221,51 +290,27 @@ static float floatVectorCmpAVX(const float* pData, int32_t numOfRows, bool isMin #if __AVX__ __m256 next; - __m256 initialVal = _mm256_loadu_ps(p); + __m256 initVal = _mm256_loadu_ps(p); p += width; if (!isMinFunc) { // max function for (int32_t i = 1; i < rounds; ++i) { next = _mm256_loadu_ps(p); - initialVal = _mm256_max_ps(initialVal, next); + initVal = _mm256_max_ps(initVal, next); p += width; } - // let sum up the final results - const float* q = (const float*)&initialVal; - v = TMAX(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMAX(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + start]) { - v = p[j + start]; - } - } + const float* q = (const float*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) } else { // min function for (int32_t i = 1; i < rounds; ++i) { next = _mm256_loadu_ps(p); - initialVal = _mm256_min_ps(initialVal, next); + initVal = _mm256_min_ps(initVal, next); p += width; } - // let sum up the final results - const float* q = (const float*)&initialVal; - v = TMIN(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMIN(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + start]) { - v = p[j + start]; - } - } + const float* q = (const float*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) } #endif @@ -282,366 +327,207 @@ static double doubleVectorCmpAVX(const double* pData, int32_t numOfRows, bool is #if __AVX__ __m256d next; - __m256d initialVal = _mm256_loadu_pd(p); + __m256d initVal = _mm256_loadu_pd(p); p += width; if (!isMinFunc) { // max function for (int32_t i = 1; i < rounds; ++i) { next = _mm256_loadu_pd(p); - initialVal = _mm256_max_pd(initialVal, next); + initVal = _mm256_max_pd(initVal, next); p += width; } // let sum up the final results - const double* q = (const double*)&initialVal; - v = TMAX(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMAX(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v < p[j + start]) { - v = p[j + start]; - } - } + const double* q = (const double*)&initVal; + EXTRACT_MAX_VAL(q, p, width, remain, v) } else { // min function for (int32_t i = 1; i < rounds; ++i) { next = _mm256_loadu_pd(p); - initialVal = _mm256_min_pd(initialVal, next); + initVal = _mm256_min_pd(initVal, next); p += width; } // let sum up the final results - const double* q = (const double*)&initialVal; - v = TMIN(q[0], q[1]); - for (int32_t k = 1; k < width; ++k) { - v = TMIN(v, q[k]); - } - - // calculate the front and the reminder items in array list - int32_t start = rounds * width; - for (int32_t j = 0; j < remain; ++j) { - if (v > p[j + start]) { - v = p[j + start]; - } - } + const double* q = (const double*)&initVal; + EXTRACT_MIN_VAL(q, p, width, remain, v) } #endif return v; } -static int32_t findFirstVal(const SColumnInfoData* pCol, int32_t start, int32_t numOfRows) { +static int32_t findFirstValPosition(const SColumnInfoData* pCol, int32_t start, int32_t numOfRows) { int32_t i = start; - while (i < (start + numOfRows)) { - if (!colDataIsNull_f(pCol->nullbitmap, i)) { - break; - } + + while (i < (start + numOfRows) && (colDataIsNull_f(pCol->nullbitmap, i) == true)) { i += 1; } return i; } -static int32_t handleInt8Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, - SMinmaxResInfo* pBuf, bool isMinFunc) { - int8_t* pData = (int8_t*)pCol->pData; - int8_t* val = (int8_t*)&pBuf->v; - - int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows <= 32 || pCtx->subsidiaries.num > 0) { - int32_t i = findFirstVal(pCol, start, numOfRows); - - if ((i < (start + numOfRows)) && (!pBuf->assign)) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - numOfElems += 1; +static void handleInt8Col(const void* data, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc, + bool signVal) { + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + pBuf->v = i8VectorCmpAVX2(data, numOfRows, isMinFunc, signVal); + } else { + if (!pBuf->assign) { + pBuf->v = ((int8_t*)data)[0]; } - if (isMinFunc) { // min - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } + if (signVal) { + const int8_t* p = (const int8_t*)data; + int8_t* v = (int8_t*)&pBuf->v; - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } - - } else { // max function - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } - - } - } else { // not has null value - // AVX2 version to speedup the loop - if (tsAVX2Enable && tsSIMDEnable) { - *val = i8VectorCmpAVX2(pData, numOfRows, isMinFunc); } else { - if (!pBuf->assign) { - *val = pData[0]; - pBuf->assign = true; - } + const uint8_t* p = (const uint8_t*)data; + uint8_t* v = (uint8_t*)&pBuf->v; - if (isMinFunc) { // min - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val > pData[i]) { - *val = pData[i]; - } - } - } else { // max - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val < pData[i]) { - *val = pData[i]; - } - } + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } } - - numOfElems = numOfRows; } - return numOfElems; + pBuf->assign = true; } -static int32_t handleInt16Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, - SMinmaxResInfo* pBuf, bool isMinFunc) { - int16_t* pData = (int16_t*)pCol->pData; - int16_t* val = (int16_t*)&pBuf->v; - - int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows <= 16 || pCtx->subsidiaries.num > 0) { - int32_t i = findFirstVal(pCol, start, numOfRows); - - if ((i < (start + numOfRows)) && (!pBuf->assign)) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - numOfElems += 1; +static void handleInt16Col(const void* data, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc, + bool signVal) { + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + pBuf->v = i16VectorCmpAVX2(data, numOfRows, isMinFunc, signVal); + } else { + if (!pBuf->assign) { + pBuf->v = ((int16_t*)data)[0]; } - if (isMinFunc) { // min - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } + if (signVal) { + const int16_t* p = (const int16_t*)data; + int16_t* v = (int16_t*)&pBuf->v; - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } - - } else { // max function - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } - - } - } else { // not has null value - // AVX2 version to speedup the loop - if (tsAVX2Enable && tsSIMDEnable) { - *val = i16VectorCmpAVX2(pData, numOfRows, isMinFunc); } else { - if (!pBuf->assign) { - *val = pData[0]; - pBuf->assign = true; - } + const uint16_t* p = (const uint16_t*)data; + uint16_t* v = (uint16_t*)&pBuf->v; - if (isMinFunc) { // min - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val > pData[i]) { - *val = pData[i]; - } - } - } else { // max - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val < pData[i]) { - *val = pData[i]; - } - } + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } } - - numOfElems = numOfRows; } - return numOfElems; + pBuf->assign = true; } -static int32_t handleInt32Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, - SMinmaxResInfo* pBuf, bool isMinFunc) { - int32_t* pData = (int32_t*)pCol->pData; - int32_t* val = (int32_t*)&pBuf->v; - - int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows <= 8 || pCtx->subsidiaries.num > 0) { - int32_t i = findFirstVal(pCol, start, numOfRows); - - if ((i < (start + numOfRows)) && (!pBuf->assign)) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - numOfElems += 1; +static void handleInt32Col(const void* data, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc, + bool signVal) { + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + pBuf->v = i32VectorCmpAVX2(data, numOfRows, isMinFunc, signVal); + } else { + if (!pBuf->assign) { + pBuf->v = ((int32_t*)data)[0]; } - if (isMinFunc) { // min - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } + if (signVal) { + const int32_t* p = (const int32_t*)data; + int32_t* v = (int32_t*)&pBuf->v; - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } - - } else { // max function - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } - - } - } else { // not has null value - // AVX2 version to speedup the loop - if (tsAVX2Enable && tsSIMDEnable) { - *val = i32VectorCmpAVX2(pData, numOfRows, isMinFunc); } else { - if (!pBuf->assign) { - *val = pData[0]; - pBuf->assign = true; - } + const uint32_t* p = (const uint32_t*)data; + uint32_t* v = (uint32_t*)&pBuf->v; - if (isMinFunc) { // min - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val > pData[i]) { - *val = pData[i]; - } - } - } else { // max - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val < pData[i]) { - *val = pData[i]; - } - } + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } } - - numOfElems = numOfRows; } - return numOfElems; + pBuf->assign = true; } -static int32_t handleInt64Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, - SMinmaxResInfo* pBuf, bool isMinFunc) { - int32_t* pData = (int32_t*)pCol->pData; - int32_t* val = (int32_t*)&pBuf->v; +static void handleInt64Col(const void* data, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc, + bool signVal) { + if (!pBuf->assign) { + pBuf->v = ((int64_t*)data)[0]; + } - int32_t numOfElems = 0; - if (pCol->hasNull || pCtx->subsidiaries.num > 0) { - int32_t i = findFirstVal(pCol, start, numOfRows); + if (signVal) { + const int64_t* p = (const int64_t*)data; + int64_t* v = &pBuf->v; - if ((i < (start + numOfRows)) && (!pBuf->assign)) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - numOfElems += 1; + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } + } else { + const uint64_t* p = (const uint64_t*)data; + uint64_t* v = (uint64_t*)&pBuf->v; - if (isMinFunc) { // min - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } - - } else { // max function - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } + if (isMinFunc) { + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *v, p); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *v, p); } - } else { // not has null value - // AVX2 version to speedup the loop + } +} + +static void handleUint8Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, + bool isMinFunc) { + const uint8_t* pData = (uint8_t*)pCol->pData; + uint8_t* val = (uint8_t*)&pBuf->v; + + // AVX2 version to speedup the loop + if (tsAVX2Enable && tsSIMDEnable) { + *val = i8VectorCmpAVX2(pData, numOfRows, isMinFunc, false); + } else { + if (!pBuf->assign) { + *val = pData[0]; + } + + if (isMinFunc) { // min + __COMPARE_EXTRACT_MIN(start, start + numOfRows, *val, pData); + } else { + __COMPARE_EXTRACT_MAX(start, start + numOfRows, *val, pData); + } + } + + pBuf->assign = true; +} + +static void handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc) { + float* pData = (float*)pCol->pData; + float* val = (float*)&pBuf->v; + + // AVX version to speedup the loop + if (tsAVXEnable && tsSIMDEnable) { + *val = floatVectorCmpAVX(pData, numOfRows, isMinFunc); + } else { if (!pBuf->assign) { *val = pData[0]; - pBuf->assign = true; } if (isMinFunc) { // min @@ -657,169 +543,39 @@ static int32_t handleInt64Col(SColumnInfoData* pCol, int32_t start, int32_t numO } } } - - numOfElems = numOfRows; - } - return numOfElems; -} - -static int32_t handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, - SMinmaxResInfo* pBuf, bool isMinFunc) { - float* pData = (float*)pCol->pData; - float* val = (float*)&pBuf->v; - - int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows < 8 || pCtx->subsidiaries.num > 0) { - int32_t i = findFirstVal(pCol, start, numOfRows); - if ((i < (start + numOfRows)) && (!pBuf->assign)) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - numOfElems += 1; - } - - if (isMinFunc) { // min - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } - } else { // max function - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } - } - } else { // not has null value - // AVX version to speedup the loop - if (tsAVXEnable && tsSIMDEnable) { - *val = (double) floatVectorCmpAVX(pData, numOfRows, isMinFunc); - } else { - if (!pBuf->assign) { - *val = pData[0]; - pBuf->assign = true; - } - - if (isMinFunc) { // min - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val > pData[i]) { - *val = pData[i]; - } - } - } else { // max - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val < pData[i]) { - *val = pData[i]; - } - } - } - } - - numOfElems = numOfRows; } - return numOfElems; + pBuf->assign = true; } -static int32_t handleDoubleCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SqlFunctionCtx* pCtx, - SMinmaxResInfo* pBuf, bool isMinFunc) { +static void handleDoubleCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc) { double* pData = (double*)pCol->pData; double* val = (double*)&pBuf->v; - int32_t numOfElems = 0; - if (pCol->hasNull || numOfRows < 4 || pCtx->subsidiaries.num > 0) { - int32_t i = findFirstVal(pCol, start, numOfRows); - - if ((i < (start + numOfRows)) && (!pBuf->assign)) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - numOfElems += 1; + // AVX version to speedup the loop + if (tsAVXEnable && tsSIMDEnable) { + *val = (double)doubleVectorCmpAVX(pData, numOfRows, isMinFunc); + } else { + if (!pBuf->assign) { + *val = pData[0]; } if (isMinFunc) { // min - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - + for (int32_t i = start; i < start + numOfRows; ++i) { if (*val > pData[i]) { *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } } - numOfElems += 1; } - } else { // max function - for (; i < start + numOfRows; ++i) { - if (colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. + } else { // max + for (int32_t i = start; i < start + numOfRows; ++i) { if (*val < pData[i]) { *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - numOfElems += 1; - } - } - } else { // not has null value - // AVX version to speedup the loop - if (tsAVXEnable && tsSIMDEnable) { - *val = (double) doubleVectorCmpAVX(pData, numOfRows, isMinFunc); - } else { - if (!pBuf->assign) { - *val = pData[0]; - pBuf->assign = true; - } - - if (isMinFunc) { // min - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val > pData[i]) { - *val = pData[i]; - } - } - } else { // max - for (int32_t i = start; i < start + numOfRows; ++i) { - if (*val < pData[i]) { - *val = pData[i]; - } } } } - - numOfElems = numOfRows; } - return numOfElems; + pBuf->assign = true; } static int32_t findRowIndex(int32_t start, int32_t num, SColumnInfoData* pCol, const char* tval) { @@ -835,6 +591,137 @@ static int32_t findRowIndex(int32_t start, int32_t num, SColumnInfoData* pCol, c return -1; } +static void doExtractVal(SColumnInfoData* pCol, int32_t i, int32_t end, SqlFunctionCtx* pCtx, SMinmaxResInfo* pBuf, + bool isMinFunc) { + if (isMinFunc) { + switch (pCol->info.type) { + case TSDB_DATA_TYPE_BOOL: + case TSDB_DATA_TYPE_TINYINT: { + const int8_t* pData = (const int8_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(int8_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_SMALLINT: { + const int16_t* pData = (const int16_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(int16_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_INT: { + const int32_t* pData = (const int32_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(int32_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_BIGINT: { + const int64_t* pData = (const int64_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, (pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_UTINYINT: { + const uint8_t* pData = (const uint8_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(uint8_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_USMALLINT: { + const uint16_t* pData = (const uint16_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(uint16_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_UINT: { + const uint32_t* pData = (const uint32_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(uint32_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_UBIGINT: { + const uint64_t* pData = (const uint64_t*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(uint64_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_FLOAT: { + const float* pData = (const float*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(float*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_DOUBLE: { + const double* pData = (const double*)pCol->pData; + __COMPARE_ACQUIRED_MIN(i, end, pCol->nullbitmap, pData, pCtx, *(double*)&(pBuf->v), &pBuf->tuplePos) + break; + } + } + } else { + switch (pCol->info.type) { + case TSDB_DATA_TYPE_BOOL: + case TSDB_DATA_TYPE_TINYINT: { + const int8_t* pData = (const int8_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(int8_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_SMALLINT: { + const int16_t* pData = (const int16_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(int16_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_INT: { + const int16_t* pData = (const int16_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(int32_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_BIGINT: { + const int64_t* pData = (const int64_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, (pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_UTINYINT: { + const uint8_t* pData = (const uint8_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(uint8_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_USMALLINT: { + const uint16_t* pData = (const uint16_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(uint16_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_UINT: { + const uint32_t* pData = (const uint32_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(uint32_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_UBIGINT: { + const uint64_t* pData = (const uint64_t*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(uint64_t*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_FLOAT: { + const float* pData = (const float*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(float*)&(pBuf->v), &pBuf->tuplePos) + break; + } + + case TSDB_DATA_TYPE_DOUBLE: { + const double* pData = (const double*)pCol->pData; + __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(double*)&(pBuf->v), &pBuf->tuplePos) + break; + } + } + } +} + int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { int32_t numOfElems = 0; @@ -857,6 +744,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { if (pInput->colDataSMAIsSet) { numOfElems = pInput->numOfRows - pAgg->numOfNull; ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); + if (numOfElems == 0) { return numOfElems; } @@ -945,212 +833,80 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { int32_t start = pInput->startRowIndex; int32_t numOfRows = pInput->numOfRows; + int32_t end = start + numOfRows; - if (IS_SIGNED_NUMERIC_TYPE(type) || type == TSDB_DATA_TYPE_BOOL) { - if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) { - numOfElems = handleInt8Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); - } else if (type == TSDB_DATA_TYPE_SMALLINT) { - numOfElems = handleInt16Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); - } else if (type == TSDB_DATA_TYPE_INT) { - numOfElems = handleInt32Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); - } else if (type == TSDB_DATA_TYPE_BIGINT) { - numOfElems = handleInt64Col(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); + if (pCol->hasNull || numOfRows < 32 || pCtx->subsidiaries.num > 0) { + int32_t i = findFirstValPosition(pCol, start, numOfRows); + + if ((i < end) && (!pBuf->assign)) { + memcpy(&pBuf->v, pCol->pData + (pCol->info.bytes * i), pCol->info.bytes); + + if (pCtx->subsidiaries.num > 0) { + pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); + } + pBuf->assign = true; + numOfElems = 1; } - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - if (type == TSDB_DATA_TYPE_UTINYINT) { - uint8_t* pData = (uint8_t*)pCol->pData; - uint8_t* val = (uint8_t*)&pBuf->v; - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } + if (i >= end) { + ASSERT(numOfElems == 0); + return numOfElems; + } - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } + doExtractVal(pCol, i, end, pCtx, pBuf, isMinFunc); + } else { + numOfElems = numOfRows; - numOfElems += 1; + switch(pCol->info.type) { + case TSDB_DATA_TYPE_BOOL: + case TSDB_DATA_TYPE_TINYINT: { + handleInt8Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, true); + break; } - } else if (type == TSDB_DATA_TYPE_USMALLINT) { - uint16_t* pData = (uint16_t*)pCol->pData; - uint16_t* val = (uint16_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; + case TSDB_DATA_TYPE_SMALLINT: { + handleInt16Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, true); + break; } - } else if (type == TSDB_DATA_TYPE_UINT) { - uint32_t* pData = (uint32_t*)pCol->pData; - uint32_t* val = (uint32_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; + case TSDB_DATA_TYPE_INT: { + handleInt32Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, true); + break; } - } else if (type == TSDB_DATA_TYPE_UBIGINT) { - uint64_t* pData = (uint64_t*)pCol->pData; - uint64_t* val = (uint64_t*)&pBuf->v; - - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; + case TSDB_DATA_TYPE_BIGINT: { + handleInt64Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, true); + break; + } + case TSDB_DATA_TYPE_UTINYINT: { + handleInt8Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); + break; + } + case TSDB_DATA_TYPE_USMALLINT: { + handleInt16Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); + break; + } + case TSDB_DATA_TYPE_UINT: { + handleInt16Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); + break; + } + case TSDB_DATA_TYPE_UBIGINT: { + handleInt16Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); + break; + } + case TSDB_DATA_TYPE_FLOAT: { + handleFloatCol(pCol, start, numOfRows, pBuf, isMinFunc); + break; + } + case TSDB_DATA_TYPE_DOUBLE: { + handleDoubleCol(pCol, start, numOfRows, pBuf, isMinFunc); + break; } } - } else if (type == TSDB_DATA_TYPE_DOUBLE) { - double* pData = (double*)pCol->pData; - double* val = (double*)&pBuf->v; - for (int32_t i = start; i < start + numOfRows; ++i) { - if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { - continue; - } - - if (!pBuf->assign) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); - } - pBuf->assign = true; - } else { - // ignore the equivalent data value - // NOTE: An faster version to avoid one additional comparison with FPU. - if (isMinFunc) { // min - if (*val > pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } else { // max - if (*val < pData[i]) { - *val = pData[i]; - if (pCtx->subsidiaries.num > 0) { - updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); - } - } - } - } - - numOfElems += 1; + _over: + if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { + pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); + pBuf->nullTupleSaved = true; } - } else if (type == TSDB_DATA_TYPE_FLOAT) { - numOfElems = handleFloatCol(pCol, start, numOfRows, pCtx, pBuf, isMinFunc); } -_over: - if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { - pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); - pBuf->nullTupleSaved = true; - } return numOfElems; } \ No newline at end of file From 969331c49e337a3f346c82a2f1c47c6cd56b5267 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 25 Nov 2022 11:14:17 +0800 Subject: [PATCH 46/83] refactor: do some internal refactor. --- source/libs/function/src/detail/tminmax.c | 41 +++++------------------ 1 file changed, 9 insertions(+), 32 deletions(-) diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index 03438350ed..e07cd38b65 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -495,29 +495,6 @@ static void handleInt64Col(const void* data, int32_t start, int32_t numOfRows, S } } -static void handleUint8Col(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, - bool isMinFunc) { - const uint8_t* pData = (uint8_t*)pCol->pData; - uint8_t* val = (uint8_t*)&pBuf->v; - - // AVX2 version to speedup the loop - if (tsAVX2Enable && tsSIMDEnable) { - *val = i8VectorCmpAVX2(pData, numOfRows, isMinFunc, false); - } else { - if (!pBuf->assign) { - *val = pData[0]; - } - - if (isMinFunc) { // min - __COMPARE_EXTRACT_MIN(start, start + numOfRows, *val, pData); - } else { - __COMPARE_EXTRACT_MAX(start, start + numOfRows, *val, pData); - } - } - - pBuf->assign = true; -} - static void handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc) { float* pData = (float*)pCol->pData; float* val = (float*)&pBuf->v; @@ -746,7 +723,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); if (numOfElems == 0) { - return numOfElems; + goto _over; } void* tval = NULL; @@ -857,7 +834,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { } else { numOfElems = numOfRows; - switch(pCol->info.type) { + switch (pCol->info.type) { case TSDB_DATA_TYPE_BOOL: case TSDB_DATA_TYPE_TINYINT: { handleInt8Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, true); @@ -884,11 +861,11 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { break; } case TSDB_DATA_TYPE_UINT: { - handleInt16Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); + handleInt32Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); break; } case TSDB_DATA_TYPE_UBIGINT: { - handleInt16Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); + handleInt64Col(pCol->pData, start, numOfRows, pBuf, isMinFunc, false); break; } case TSDB_DATA_TYPE_FLOAT: { @@ -900,12 +877,12 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { break; } } + } - _over: - if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { - pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); - pBuf->nullTupleSaved = true; - } +_over: + if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { + pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); + pBuf->nullTupleSaved = true; } return numOfElems; From f8bc4df4afe96690789e279009f57845aca0462e Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 25 Nov 2022 11:58:46 +0800 Subject: [PATCH 47/83] refactor: do some internal refactor. --- source/dnode/vnode/inc/vnode.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/source/dnode/vnode/inc/vnode.h b/source/dnode/vnode/inc/vnode.h index 149c3fbaa9..e25c899036 100644 --- a/source/dnode/vnode/inc/vnode.h +++ b/source/dnode/vnode/inc/vnode.h @@ -108,9 +108,11 @@ int metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName); int metaGetTableUidByName(void *meta, char *tbName, uint64_t *uid); int metaGetTableTypeByName(void *meta, char *tbName, ETableType *tbType); bool metaIsTableExist(SMeta *pMeta, tb_uid_t uid); -int32_t metaGetCachedTableUidList(SMeta *pMeta, tb_uid_t suid, const uint8_t *key, int32_t keyLen, SArray *pList, bool* acquired); +int32_t metaGetCachedTableUidList(SMeta *pMeta, tb_uid_t suid, const uint8_t *key, int32_t keyLen, SArray *pList, + bool *acquired); int32_t metaUidFilterCachePut(SMeta *pMeta, uint64_t suid, const void *pKey, int32_t keyLen, void *pPayload, int32_t payloadLen, double selectivityRatio); +int32_t metaUidCacheClear(SMeta *pMeta, uint64_t suid); typedef struct SMetaFltParam { tb_uid_t suid; From 59a15719451eea1e32b23ee1e7852c20569a5d7b Mon Sep 17 00:00:00 2001 From: Minglei Jin Date: Fri, 25 Nov 2022 18:56:00 +0800 Subject: [PATCH 48/83] fix(meta/cache): invalidate suid cache when create/drop ctb & update tag --- source/dnode/vnode/src/meta/metaTable.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/source/dnode/vnode/src/meta/metaTable.c b/source/dnode/vnode/src/meta/metaTable.c index 4d0a0a50a4..dff8adf29b 100644 --- a/source/dnode/vnode/src/meta/metaTable.c +++ b/source/dnode/vnode/src/meta/metaTable.c @@ -458,6 +458,7 @@ int metaCreateTable(SMeta *pMeta, int64_t version, SVCreateTbReq *pReq, STableMe metaWLock(pMeta); metaUpdateStbStats(pMeta, me.ctbEntry.suid, 1); + metaUidCacheClear(pMeta, me.ctbEntry.suid); metaULock(pMeta); } else { me.ntbEntry.ctime = pReq->ctime; @@ -681,6 +682,7 @@ static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) { --pMeta->pVnode->config.vndStats.numOfCTables; metaUpdateStbStats(pMeta, e.ctbEntry.suid, -1); + metaUidCacheClear(pMeta, e.ctbEntry.suid); } else if (e.type == TSDB_NORMAL_TABLE) { // drop schema.db (todo) @@ -691,6 +693,7 @@ static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) { // drop schema.db (todo) metaStatsCacheDrop(pMeta, uid); + metaUidCacheClear(pMeta, uid); --pMeta->pVnode->config.vndStats.numOfSTables; } @@ -1069,6 +1072,8 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA tdbTbUpsert(pMeta->pCtbIdx, &ctbIdxKey, sizeof(ctbIdxKey), ctbEntry.ctbEntry.pTags, ((STag *)(ctbEntry.ctbEntry.pTags))->len, &pMeta->txn); + metaUidCacheClear(pMeta, ctbEntry.ctbEntry.suid); + metaULock(pMeta); tDecoderClear(&dc1); From d77f5fa70c001a5c756edb3cec91441fec57fb53 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 25 Nov 2022 22:54:54 +0800 Subject: [PATCH 49/83] fix(query): fix error in windows and darwin system. --- include/os/os.h | 2 +- include/util/tpagedbuf.h | 7 ++----- source/os/src/osMemory.c | 4 ++++ source/util/src/tpagedbuf.c | 4 ++-- 4 files changed, 9 insertions(+), 8 deletions(-) diff --git a/include/os/os.h b/include/os/os.h index 0334cd4d95..ab4d0a406e 100644 --- a/include/os/os.h +++ b/include/os/os.h @@ -43,6 +43,7 @@ extern "C" { #include #include #include +#include #if defined(DARWIN) #else @@ -81,7 +82,6 @@ extern "C" { #include #include #include -#include #if __AVX__ #include diff --git a/include/util/tpagedbuf.h b/include/util/tpagedbuf.h index 5f81ccfcfd..73af65997d 100644 --- a/include/util/tpagedbuf.h +++ b/include/util/tpagedbuf.h @@ -24,12 +24,9 @@ extern "C" { #endif -typedef struct SArray* SIDList; typedef struct SPageInfo SPageInfo; typedef struct SDiskbasedBuf SDiskbasedBuf; -#define DEFAULT_INTERN_BUF_PAGE_SIZE (1024LL) // in bytes - typedef struct SFilePage { int32_t num; char data[]; @@ -69,7 +66,7 @@ void* getNewBufPage(SDiskbasedBuf* pBuf, int32_t* pageId); * @param pBuf * @return */ -SIDList getDataBufPagesIdList(SDiskbasedBuf* pBuf); +SArray* getDataBufPagesIdList(SDiskbasedBuf* pBuf); /** * get the specified buffer page by id @@ -111,7 +108,7 @@ void destroyDiskbasedBuf(SDiskbasedBuf* pBuf); * @param pList * @return */ -SPageInfo* getLastPageInfo(SIDList pList); +SPageInfo* getLastPageInfo(SArray* pList); /** * diff --git a/source/os/src/osMemory.c b/source/os/src/osMemory.c index 1facff1f3b..2f30e8977a 100644 --- a/source/os/src/osMemory.c +++ b/source/os/src/osMemory.c @@ -350,6 +350,10 @@ void* taosMemoryMallocAlign(uint32_t alignment, int64_t size) { #ifdef USE_TD_MEMORY ASSERT(0); #else +#if defined(LINUX) return memalign(alignment, size); +#else + return taosMemoryMalloc(size); +#endif #endif } diff --git a/source/util/src/tpagedbuf.c b/source/util/src/tpagedbuf.c index 79ea10552c..e1a43ace47 100644 --- a/source/util/src/tpagedbuf.c +++ b/source/util/src/tpagedbuf.c @@ -495,7 +495,7 @@ void releaseBufPageInfo(SDiskbasedBuf* pBuf, SPageInfo* pi) { size_t getTotalBufSize(const SDiskbasedBuf* pBuf) { return (size_t)pBuf->totalBufSize; } -SIDList getDataBufPagesIdList(SDiskbasedBuf* pBuf) { +SArray* getDataBufPagesIdList(SDiskbasedBuf* pBuf) { ASSERT(pBuf != NULL); return pBuf->pIdList; } @@ -561,7 +561,7 @@ void destroyDiskbasedBuf(SDiskbasedBuf* pBuf) { taosMemoryFreeClear(pBuf); } -SPageInfo* getLastPageInfo(SIDList pList) { +SPageInfo* getLastPageInfo(SArray* pList) { size_t size = taosArrayGetSize(pList); SPageInfo* pPgInfo = taosArrayGetP(pList, size - 1); return pPgInfo; From 86fda9f3da231bbecc4e97ddcb9d1927a29666a4 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 25 Nov 2022 23:06:32 +0800 Subject: [PATCH 50/83] fix(query): fix error in windows and darwin system. --- source/libs/executor/src/executorimpl.c | 2 +- source/libs/function/src/tpercentile.c | 2 +- source/util/test/pageBufferTest.cpp | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/source/libs/executor/src/executorimpl.c b/source/libs/executor/src/executorimpl.c index 49471dc959..9ad0aae3e4 100644 --- a/source/libs/executor/src/executorimpl.c +++ b/source/libs/executor/src/executorimpl.c @@ -244,7 +244,7 @@ static int32_t addNewWindowResultBuf(SResultRow* pWindowRes, SDiskbasedBuf* pRes // in the first scan, new space needed for results int32_t pageId = -1; - SIDList list = getDataBufPagesIdList(pResultBuf); + SArray* list = getDataBufPagesIdList(pResultBuf); if (taosArrayGetSize(list) == 0) { pData = getNewBufPage(pResultBuf, &pageId); diff --git a/source/libs/function/src/tpercentile.c b/source/libs/function/src/tpercentile.c index 0924106476..e5727f1472 100644 --- a/source/libs/function/src/tpercentile.c +++ b/source/libs/function/src/tpercentile.c @@ -494,7 +494,7 @@ double getPercentileImpl(tMemBucket *pMemBucket, int32_t count, double fraction) resetSlotInfo(pMemBucket); int32_t groupId = getGroupId(pMemBucket->numOfSlots, i, pMemBucket->times - 1); - SIDList list = taosHashGet(pMemBucket->groupPagesMap, &groupId, sizeof(groupId)); + SArray* list = taosHashGet(pMemBucket->groupPagesMap, &groupId, sizeof(groupId)); ASSERT(list != NULL && list->size > 0); for (int32_t f = 0; f < list->size; ++f) { diff --git a/source/util/test/pageBufferTest.cpp b/source/util/test/pageBufferTest.cpp index 869bb2a76d..00ed804930 100644 --- a/source/util/test/pageBufferTest.cpp +++ b/source/util/test/pageBufferTest.cpp @@ -23,7 +23,7 @@ void simpleTest() { ASSERT_EQ(getTotalBufSize(pBuf), 1024); - SIDList list = getDataBufPagesIdList(pBuf); + SArray* list = getDataBufPagesIdList(pBuf); ASSERT_EQ(taosArrayGetSize(list), 1); // ASSERT_EQ(getNumOfBufGroupId(pBuf), 1); From 99ea0abffa13020f7b1035c36cf1df7f2f5892c5 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 25 Nov 2022 23:09:31 +0800 Subject: [PATCH 51/83] fix(query): add return value. --- source/os/src/osSysinfo.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/source/os/src/osSysinfo.c b/source/os/src/osSysinfo.c index 93eebd2437..f336b84e1e 100644 --- a/source/os/src/osSysinfo.c +++ b/source/os/src/osSysinfo.c @@ -504,9 +504,9 @@ int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma) { // Ref to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77756 __cpuid_fix(7u, eax, ebx, ecx, edx); *avx2 = (char) ((ebx & bit_AVX2) == bit_AVX2); - return 0; - #endif + + return 0; } int32_t taosGetTotalMemory(int64_t *totalKB) { From f9d1726c890471b0f0339e560ac76189134f3434 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Fri, 25 Nov 2022 23:27:36 +0800 Subject: [PATCH 52/83] fix(query): fix error in ut. --- include/util/thash.h | 2 +- source/dnode/mgmt/test/sut/src/client.cpp | 1 + source/libs/parser/test/mockCatalogService.cpp | 2 +- source/libs/scheduler/test/schedulerTests.cpp | 1 + source/libs/transport/test/transUT.cpp | 2 ++ source/util/src/thash.c | 2 +- 6 files changed, 7 insertions(+), 3 deletions(-) diff --git a/include/util/thash.h b/include/util/thash.h index a04f78a3d1..08caad495d 100644 --- a/include/util/thash.h +++ b/include/util/thash.h @@ -213,7 +213,7 @@ void taosHashSetEqualFp(SHashObj *pHashObj, _equal_fn_t fp); */ void taosHashSetFreeFp(SHashObj *pHashObj, _hash_free_fn_t fp); -//int64_t taosHashGetCompTimes(SHashObj *pHashObj); +int64_t taosHashGetCompTimes(SHashObj *pHashObj); #ifdef __cplusplus } diff --git a/source/dnode/mgmt/test/sut/src/client.cpp b/source/dnode/mgmt/test/sut/src/client.cpp index 6b4c23c0de..a27a511651 100644 --- a/source/dnode/mgmt/test/sut/src/client.cpp +++ b/source/dnode/mgmt/test/sut/src/client.cpp @@ -15,6 +15,7 @@ #include "sut.h" #include "tdatablock.h" +#include "tmisce.h" static void processClientRsp(void* parent, SRpcMsg* pRsp, SEpSet* pEpSet) { TestClient* client = (TestClient*)parent; diff --git a/source/libs/parser/test/mockCatalogService.cpp b/source/libs/parser/test/mockCatalogService.cpp index 95f7af435d..be2e4b90b9 100644 --- a/source/libs/parser/test/mockCatalogService.cpp +++ b/source/libs/parser/test/mockCatalogService.cpp @@ -20,9 +20,9 @@ #include #include -#include "tdatablock.h" #include "tname.h" #include "ttypes.h" +#include "tmisce.h" std::unique_ptr g_mockCatalogService; diff --git a/source/libs/scheduler/test/schedulerTests.cpp b/source/libs/scheduler/test/schedulerTests.cpp index 97e14b617c..5605a4b842 100644 --- a/source/libs/scheduler/test/schedulerTests.cpp +++ b/source/libs/scheduler/test/schedulerTests.cpp @@ -38,6 +38,7 @@ #include "tglobal.h" #include "trpc.h" #include "tvariant.h" +#include "tmisce.h" #pragma GCC diagnostic push #pragma GCC diagnostic ignored "-Wwrite-strings" diff --git a/source/libs/transport/test/transUT.cpp b/source/libs/transport/test/transUT.cpp index 92f0bf11cf..88a1e2564f 100644 --- a/source/libs/transport/test/transUT.cpp +++ b/source/libs/transport/test/transUT.cpp @@ -20,6 +20,8 @@ #include "tlog.h" #include "transLog.h" #include "trpc.h" +#include "tmisce.h" + using namespace std; const char *label = "APP"; diff --git a/source/util/src/thash.c b/source/util/src/thash.c index fe6c0a8af5..a0411483ca 100644 --- a/source/util/src/thash.c +++ b/source/util/src/thash.c @@ -889,4 +889,4 @@ void *taosHashAcquire(SHashObj *pHashObj, const void *key, size_t keyLen) { void taosHashRelease(SHashObj *pHashObj, void *p) { taosHashCancelIterate(pHashObj, p); } -//int64_t taosHashGetCompTimes(SHashObj *pHashObj) { return atomic_load_64(&pHashObj->compTimes); } +int64_t taosHashGetCompTimes(SHashObj *pHashObj) { return 0 /*atomic_load_64(&pHashObj->compTimes)*/; } From c6795b68d2a98eec103a4e3d3f9d4e97648571be Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 00:12:27 +0800 Subject: [PATCH 53/83] fix(query): check list size before iterate it. --- source/dnode/vnode/src/meta/metaCache.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 05b6209c56..4aa6ed3744 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -519,20 +519,20 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int // remove the lru cache that are expired due to the tags value update, or creating, or dropping, of child tables int32_t metaUidCacheClear(SMeta* pMeta, uint64_t suid) { STagFilterResEntry* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); - if (pEntry == NULL) { + if (pEntry == NULL || listNEles(&pEntry->list) == 0) { return TSDB_CODE_SUCCESS; } - int32_t keyLen = sizeof(uint64_t) + 128; - char* p = taosMemoryMalloc(keyLen); - *(uint64_t*)p = pEntry->suid; + int32_t keyLen = sizeof(uint64_t) * 3; + uint64_t p[3] = {0}; + p[0] = suid; SListIter iter = {0}; tdListInitIter(&pEntry->list, &iter, TD_LIST_FORWARD); SListNode* pNode = NULL; while ((pNode = tdListNext(&iter)) != NULL) { - memcpy(p + sizeof(suid), pNode->data, 128); + memcpy(&p[1], pNode->data, 16); taosLRUCacheErase(pMeta->pCache->sTagFilterResCache.pUidResCache, p, keyLen); } From 802ca0f607ff7d8241ebf0765f04987f710f6bf9 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 00:58:29 +0800 Subject: [PATCH 54/83] fix(query): fix memory leak. --- source/dnode/vnode/src/meta/metaCache.c | 7 +++++++ source/libs/executor/src/executil.c | 1 + 2 files changed, 8 insertions(+) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 4aa6ed3744..45f70e1c92 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -90,6 +90,12 @@ static void statsCacheClose(SMeta* pMeta) { } } +static void freeCacheEntryFp(void* param) { + STagFilterResEntry** p = param; + tdListEmpty(&(*p)->list); + taosMemoryFreeClear(*p); +} + int32_t metaCacheOpen(SMeta* pMeta) { int32_t code = 0; SMetaCache* pCache = NULL; @@ -132,6 +138,7 @@ int32_t metaCacheOpen(SMeta* pMeta) { goto _err2; } + taosHashSetFreeFp(pCache->sTagFilterResCache.pTableEntry, freeCacheEntryFp); pMeta->pCache = pCache; return code; diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 6ee2dce59b..21fba22af4 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1017,6 +1017,7 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); + taosMemoryFree(pPayload); } else { qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)); } From 98771bb461ce1b8aaa3ee5f4973c7d96b76fb728 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 01:50:17 +0800 Subject: [PATCH 55/83] fix(query): deprecate the error reported by asan. --- source/dnode/vnode/src/meta/metaCache.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 45f70e1c92..c590f0d014 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -497,17 +497,17 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int SLRUCache* pCache = pMeta->pCache->sTagFilterResCache.pUidResCache; SHashObj* pTableEntry = pMeta->pCache->sTagFilterResCache.pTableEntry; - void* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); + STagFilterResEntry** pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); if (pEntry == NULL) { STagFilterResEntry* p = taosMemoryMalloc(sizeof(STagFilterResEntry)); p->qTimes = 0; tdListInit(&p->list, keyLen); - - pEntry = &p; - taosHashPut(pTableEntry, &suid, sizeof(uint64_t), pEntry, POINTER_BYTES); + taosHashPut(pTableEntry, &suid, sizeof(uint64_t), &p, POINTER_BYTES); + tdListAppend(&p->list, pKey); + } else { + tdListAppend(&(*pEntry)->list, pKey); } - tdListAppend(&(*(STagFilterResEntry**)pEntry)->list, pKey); uint64_t* pBuf = pMeta->pCache->sTagFilterResCache.keyBuf; pBuf[0] = suid; From 0015e65f5b2e89d8ff6b37d180c9a1f9fe615fb7 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 02:13:46 +0800 Subject: [PATCH 56/83] fix(query): remove invalid free --- source/dnode/vnode/src/meta/metaCache.c | 1 - source/libs/executor/src/executil.c | 1 - 2 files changed, 2 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index c590f0d014..fa803627ea 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -508,7 +508,6 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int tdListAppend(&(*pEntry)->list, pKey); } - uint64_t* pBuf = pMeta->pCache->sTagFilterResCache.keyBuf; pBuf[0] = suid; diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 21fba22af4..6ee2dce59b 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -1017,7 +1017,6 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); - taosMemoryFree(pPayload); } else { qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)); } From 861b9b82642f40967722e6fa91bc986a2fbbe516 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 11:46:36 +0800 Subject: [PATCH 57/83] fix(query): fix memory leak. --- source/dnode/vnode/src/meta/metaCache.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index fa803627ea..5277e25d4a 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -443,6 +443,7 @@ int32_t metaGetCachedTableUidList(SMeta* pMeta, tb_uid_t suid, const uint8_t* pK taosArrayAddBatch(pList1, p + sizeof(int32_t), size); (*pEntry)->qTimes += 1; + taosLRUCacheRelease(pCache, pHandle, false); // check if scanning all items are necessary or not if ((*pEntry)->qTimes >= 5000 && TD_DLIST_NELES(&(*pEntry)->list) > 10) { @@ -478,12 +479,20 @@ int32_t metaGetCachedTableUidList(SMeta* pMeta, tb_uid_t suid, const uint8_t* pK return TSDB_CODE_SUCCESS; } +static void freePayload(const void* key, size_t keyLen, void* value) { + if (value == NULL) { + return; + } + taosMemoryFree(value); +} + // check both the payload size and selectivity ratio int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int32_t keyLen, void* pPayload, int32_t payloadLen, double selectivityRatio) { if (selectivityRatio > tsSelectivityRatio) { metaDebug("vgId:%d, suid:%" PRIu64 " failed to add to uid list cache, due to selectivity ratio %.2f less than threshold %.2f", TD_VID(pMeta->pVnode), suid, selectivityRatio, tsSelectivityRatio); + taosMemoryFree(pPayload); return TSDB_CODE_SUCCESS; } @@ -491,13 +500,14 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int metaDebug("vgId:%d, suid:%" PRIu64 " failed to add to uid list cache, due to payload length %d greater than threshold %d", TD_VID(pMeta->pVnode), suid, payloadLen, tsTagFilterResCacheSize); + taosMemoryFree(pPayload); return TSDB_CODE_SUCCESS; } SLRUCache* pCache = pMeta->pCache->sTagFilterResCache.pUidResCache; - SHashObj* pTableEntry = pMeta->pCache->sTagFilterResCache.pTableEntry; + SHashObj* pTableEntry = pMeta->pCache->sTagFilterResCache.pTableEntry; - STagFilterResEntry** pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); + STagFilterResEntry** pEntry = taosHashGet(pTableEntry, &suid, sizeof(uint64_t)); if (pEntry == NULL) { STagFilterResEntry* p = taosMemoryMalloc(sizeof(STagFilterResEntry)); p->qTimes = 0; @@ -515,7 +525,7 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int ASSERT(sizeof(uint64_t) + keyLen == 24); // add to cache. - taosLRUCacheInsert(pCache, pBuf, sizeof(uint64_t) + keyLen, pPayload, payloadLen, NULL, NULL, TAOS_LRU_PRIORITY_LOW); + taosLRUCacheInsert(pCache, pBuf, sizeof(uint64_t) + keyLen, pPayload, payloadLen, freePayload, NULL, TAOS_LRU_PRIORITY_LOW); metaDebug("vgId:%d, suid:%"PRIu64" list cache added into cache, total:%d, tables:%d", TD_VID(pMeta->pVnode), suid, (int32_t) taosLRUCacheGetUsage(pCache), taosHashGetSize(pTableEntry)); From 60d6b8e2eec0b0b852c3b276a37aeff8545d69b0 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 11:51:15 +0800 Subject: [PATCH 58/83] fix(query): change the ptr --- source/dnode/vnode/src/meta/metaCache.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/source/dnode/vnode/src/meta/metaCache.c b/source/dnode/vnode/src/meta/metaCache.c index 5277e25d4a..6a704d0425 100644 --- a/source/dnode/vnode/src/meta/metaCache.c +++ b/source/dnode/vnode/src/meta/metaCache.c @@ -534,8 +534,8 @@ int32_t metaUidFilterCachePut(SMeta* pMeta, uint64_t suid, const void* pKey, int // remove the lru cache that are expired due to the tags value update, or creating, or dropping, of child tables int32_t metaUidCacheClear(SMeta* pMeta, uint64_t suid) { - STagFilterResEntry* pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); - if (pEntry == NULL || listNEles(&pEntry->list) == 0) { + STagFilterResEntry** pEntry = taosHashGet(pMeta->pCache->sTagFilterResCache.pTableEntry, &suid, sizeof(uint64_t)); + if (pEntry == NULL || listNEles(&(*pEntry)->list) == 0) { return TSDB_CODE_SUCCESS; } @@ -544,7 +544,7 @@ int32_t metaUidCacheClear(SMeta* pMeta, uint64_t suid) { p[0] = suid; SListIter iter = {0}; - tdListInitIter(&pEntry->list, &iter, TD_LIST_FORWARD); + tdListInitIter(&(*pEntry)->list, &iter, TD_LIST_FORWARD); SListNode* pNode = NULL; while ((pNode = tdListNext(&iter)) != NULL) { @@ -552,8 +552,8 @@ int32_t metaUidCacheClear(SMeta* pMeta, uint64_t suid) { taosLRUCacheErase(pMeta->pCache->sTagFilterResCache.pUidResCache, p, keyLen); } - pEntry->qTimes = 0; - tdListEmpty(&pEntry->list); + (*pEntry)->qTimes = 0; + tdListEmpty(&(*pEntry)->list); return TSDB_CODE_SUCCESS; } From 257137e55e814dc2988389020cd8657e418e49b4 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 12:15:49 +0800 Subject: [PATCH 59/83] fix(query): fix memory leak. --- source/libs/executor/src/executil.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 6ee2dce59b..c072a5f1aa 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -985,6 +985,8 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, tMD5Init(&context); tMD5Update(&context, (uint8_t*)payload, (uint32_t)len); tMD5Final(&context); + + taosMemoryFree(payload); } bool acquired = false; From 965ee9a6c8067dd63b4bd32c8752beb1f6e3d530 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 15:32:36 +0800 Subject: [PATCH 60/83] fix(query): keep block order in exchange operator. --- source/libs/executor/src/exchangeoperator.c | 16 +++++++++++----- source/libs/executor/src/tsort.c | 2 +- source/libs/function/src/detail/tminmax.c | 2 +- source/util/src/tarray.c | 2 ++ tests/script/tsim/parser/select_with_tags.sim | 3 +++ 5 files changed, 18 insertions(+), 7 deletions(-) diff --git a/source/libs/executor/src/exchangeoperator.c b/source/libs/executor/src/exchangeoperator.c index 52aa3db0fd..0efdeca5b6 100644 --- a/source/libs/executor/src/exchangeoperator.c +++ b/source/libs/executor/src/exchangeoperator.c @@ -182,10 +182,15 @@ static SSDataBlock* doLoadRemoteDataImpl(SOperatorInfo* pOperator) { } // we have buffered retrieved datablock, return it directly - SSDataBlock** p = taosArrayPop(pExchangeInfo->pResultBlockList); + SSDataBlock* p = NULL; + if (taosArrayGetSize(pExchangeInfo->pResultBlockList) > 0) { + p = taosArrayGetP(pExchangeInfo->pResultBlockList, 0); + taosArrayRemove(pExchangeInfo->pResultBlockList, 0); + } + if (p != NULL) { - taosArrayPush(pExchangeInfo->pRecycledBlocks, p); - return *p; + taosArrayPush(pExchangeInfo->pRecycledBlocks, &p); + return p; } else { if (pExchangeInfo->seqLoadData) { seqLoadRemoteData(pOperator); @@ -196,9 +201,10 @@ static SSDataBlock* doLoadRemoteDataImpl(SOperatorInfo* pOperator) { if (taosArrayGetSize(pExchangeInfo->pResultBlockList) == 0) { return NULL; } else { - p = taosArrayPop(pExchangeInfo->pResultBlockList); + p = taosArrayGetP(pExchangeInfo->pResultBlockList, 0); + taosArrayRemove(pExchangeInfo->pResultBlockList, 0); taosArrayPush(pExchangeInfo->pRecycledBlocks, p); - return *p; + return p; } } } diff --git a/source/libs/executor/src/tsort.c b/source/libs/executor/src/tsort.c index 02f2b15a8f..e0a0b9442e 100644 --- a/source/libs/executor/src/tsort.c +++ b/source/libs/executor/src/tsort.c @@ -801,7 +801,7 @@ STupleHandle* tsortNextTuple(SSortHandle* pHandle) { index = tMergeTreeGetChosenIndex(pHandle->pMergeTree); pSource = pHandle->cmpParam.pSources[index]; - assert(pSource->src.pBlock != NULL); + ASSERT(pSource->src.pBlock != NULL); pHandle->tupleHandle.rowIndex = pSource->src.rowIndex; pHandle->tupleHandle.pBlock = pSource->src.pBlock; diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index e07cd38b65..46e68f46ec 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -649,7 +649,7 @@ static void doExtractVal(SColumnInfoData* pCol, int32_t i, int32_t end, SqlFunct } case TSDB_DATA_TYPE_INT: { - const int16_t* pData = (const int16_t*)pCol->pData; + const int32_t* pData = (const int32_t*)pCol->pData; __COMPARE_ACQUIRED_MAX(i, end, pCol->nullbitmap, pData, pCtx, *(int32_t*)&(pBuf->v), &pBuf->tuplePos) break; } diff --git a/source/util/src/tarray.c b/source/util/src/tarray.c index 5703d8f8f4..bcf9a77950 100644 --- a/source/util/src/tarray.c +++ b/source/util/src/tarray.c @@ -17,6 +17,8 @@ #include "tarray.h" #include "tcoding.h" +// todo refactor API + SArray* taosArrayInit(size_t size, size_t elemSize) { assert(elemSize > 0); diff --git a/tests/script/tsim/parser/select_with_tags.sim b/tests/script/tsim/parser/select_with_tags.sim index 62f1771f03..0e777de7e8 100644 --- a/tests/script/tsim/parser/select_with_tags.sim +++ b/tests/script/tsim/parser/select_with_tags.sim @@ -396,7 +396,10 @@ if $row != 12800 then return -1 endi +print $data00 , $data01 + if $data00 != @select_tags_tb0@ then + print expect select_tags_tb0 , actual: $data00 return -1 endi From 9997ff9adaf5f8fe76de4ea3b62ac8fca6b56dd5 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 15:33:53 +0800 Subject: [PATCH 61/83] fix(query): fix one typo --- source/libs/executor/src/exchangeoperator.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/libs/executor/src/exchangeoperator.c b/source/libs/executor/src/exchangeoperator.c index 0efdeca5b6..de918e0034 100644 --- a/source/libs/executor/src/exchangeoperator.c +++ b/source/libs/executor/src/exchangeoperator.c @@ -203,7 +203,7 @@ static SSDataBlock* doLoadRemoteDataImpl(SOperatorInfo* pOperator) { } else { p = taosArrayGetP(pExchangeInfo->pResultBlockList, 0); taosArrayRemove(pExchangeInfo->pResultBlockList, 0); - taosArrayPush(pExchangeInfo->pRecycledBlocks, p); + taosArrayPush(pExchangeInfo->pRecycledBlocks, &p); return p; } } From 0948216ca90755b958e4d6181b19320488d66c30 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sat, 26 Nov 2022 21:13:34 +0800 Subject: [PATCH 62/83] fix(query): fix dead lock. --- source/dnode/vnode/inc/vnode.h | 1 + source/dnode/vnode/src/meta/metaQuery.c | 6 ++++-- source/dnode/vnode/src/meta/metaSma.c | 2 +- source/dnode/vnode/src/sma/smaRollup.c | 4 ++-- source/dnode/vnode/src/tq/tqExec.c | 2 +- source/dnode/vnode/src/tq/tqRead.c | 2 +- source/dnode/vnode/src/tsdb/tsdbCache.c | 2 +- source/dnode/vnode/src/tsdb/tsdbRead.c | 4 ++-- source/libs/executor/src/executil.c | 4 ++-- source/libs/executor/src/executor.c | 2 +- source/libs/executor/src/executorimpl.c | 4 ++-- source/libs/executor/src/scanoperator.c | 4 ++-- source/libs/executor/src/sysscanoperator.c | 7 +++++++ 13 files changed, 27 insertions(+), 17 deletions(-) diff --git a/source/dnode/vnode/inc/vnode.h b/source/dnode/vnode/inc/vnode.h index e25c899036..7ef3207b4d 100644 --- a/source/dnode/vnode/inc/vnode.h +++ b/source/dnode/vnode/inc/vnode.h @@ -99,6 +99,7 @@ void metaReaderInit(SMetaReader *pReader, SMeta *pMeta, int32_t flags); void metaReaderReleaseLock(SMetaReader *pReader); void metaReaderClear(SMetaReader *pReader); int32_t metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid); +int32_t metaGetTableEntryByUidCache(SMetaReader *pReader, tb_uid_t uid); int metaGetTableEntryByName(SMetaReader *pReader, const char *name); int32_t metaGetTableTags(SMeta *pMeta, uint64_t suid, SArray *uidList, SHashObj *tags); int32_t metaGetTableTagsByUids(SMeta *pMeta, int64_t suid, SArray *uidList, SHashObj *tags); diff --git a/source/dnode/vnode/src/meta/metaQuery.c b/source/dnode/vnode/src/meta/metaQuery.c index 0f409eecd9..0257aede3d 100644 --- a/source/dnode/vnode/src/meta/metaQuery.c +++ b/source/dnode/vnode/src/meta/metaQuery.c @@ -153,7 +153,6 @@ bool metaIsTableExist(SMeta *pMeta, tb_uid_t uid) { int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) { SMeta *pMeta = pReader->pMeta; - /* int64_t version1; // query uid.idx @@ -164,7 +163,10 @@ int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) { version1 = ((SUidIdxVal *)pReader->pBuf)[0].version; return metaGetTableEntryByVersion(pReader, version1, uid); - */ +} + +int metaGetTableEntryByUidCache(SMetaReader *pReader, tb_uid_t uid) { + SMeta *pMeta = pReader->pMeta; SMetaInfo info; if (metaGetInfo(pMeta, uid, &info, pReader) == TSDB_CODE_NOT_FOUND) { diff --git a/source/dnode/vnode/src/meta/metaSma.c b/source/dnode/vnode/src/meta/metaSma.c index 3ada7d1814..52452bf710 100644 --- a/source/dnode/vnode/src/meta/metaSma.c +++ b/source/dnode/vnode/src/meta/metaSma.c @@ -36,7 +36,7 @@ int32_t metaCreateTSma(SMeta *pMeta, int64_t version, SSmaCfg *pCfg) { // validate req // save smaIndex metaReaderInit(&mr, pMeta, 0); - if (metaGetTableEntryByUid(&mr, pCfg->indexUid) == 0) { + if (metaGetTableEntryByUidCache(&mr, pCfg->indexUid) == 0) { #if 1 terrno = TSDB_CODE_TSMA_ALREADY_EXIST; metaReaderClear(&mr); diff --git a/source/dnode/vnode/src/sma/smaRollup.c b/source/dnode/vnode/src/sma/smaRollup.c index 75fb566438..c5f040c987 100644 --- a/source/dnode/vnode/src/sma/smaRollup.c +++ b/source/dnode/vnode/src/sma/smaRollup.c @@ -921,7 +921,7 @@ static int32_t tdRSmaInfoClone(SSma *pSma, SRSmaInfo *pInfo) { SMetaReader mr = {0}; metaReaderInit(&mr, SMA_META(pSma), 0); smaDebug("vgId:%d, rsma clone qTaskInfo for suid:%" PRIi64, SMA_VID(pSma), pInfo->suid); - if (metaGetTableEntryByUid(&mr, pInfo->suid) < 0) { + if (metaGetTableEntryByUidCache(&mr, pInfo->suid) < 0) { smaError("vgId:%d, rsma clone, failed to get table meta for %" PRIi64 " since %s", SMA_VID(pSma), pInfo->suid, terrstr()); goto _err; @@ -1125,7 +1125,7 @@ static int32_t tdRSmaRestoreQTaskInfoInit(SSma *pSma, int64_t *nTables) { for (int64_t i = 0; i < arrSize; ++i) { tb_uid_t suid = *(tb_uid_t *)taosArrayGet(suidList, i); smaDebug("vgId:%d, rsma restore, suid is %" PRIi64, TD_VID(pVnode), suid); - if (metaGetTableEntryByUid(&mr, suid) < 0) { + if (metaGetTableEntryByUidCache(&mr, suid) < 0) { smaError("vgId:%d, rsma restore, failed to get table meta for %" PRIi64 " since %s", TD_VID(pVnode), suid, terrstr()); goto _err; diff --git a/source/dnode/vnode/src/tq/tqExec.c b/source/dnode/vnode/src/tq/tqExec.c index 3887f72740..8bcfe53712 100644 --- a/source/dnode/vnode/src/tq/tqExec.c +++ b/source/dnode/vnode/src/tq/tqExec.c @@ -48,7 +48,7 @@ static int32_t tqAddTbNameToRsp(const STQ* pTq, int64_t uid, SMqDataRsp* pRsp, i SMetaReader mr = {0}; metaReaderInit(&mr, pTq->pVnode->pMeta, 0); // TODO add reference to gurantee success - if (metaGetTableEntryByUid(&mr, uid) < 0) { + if (metaGetTableEntryByUidCache(&mr, uid) < 0) { metaReaderClear(&mr); return -1; } diff --git a/source/dnode/vnode/src/tq/tqRead.c b/source/dnode/vnode/src/tq/tqRead.c index afb7ac39de..e41b1d8aa8 100644 --- a/source/dnode/vnode/src/tq/tqRead.c +++ b/source/dnode/vnode/src/tq/tqRead.c @@ -766,7 +766,7 @@ int32_t tqUpdateTbUidList(STQ* pTq, const SArray* tbUidList, bool isAdd) { for (int32_t i = 0; i < taosArrayGetSize(tbUidList); ++i) { uint64_t* id = (uint64_t*)taosArrayGet(tbUidList, i); - int32_t code = metaGetTableEntryByUid(&mr, *id); + int32_t code = metaGetTableEntryByUidCache(&mr, *id); if (code != TSDB_CODE_SUCCESS) { qError("failed to get table meta, uid:%" PRIu64 " code:%s", *id, tstrerror(terrno)); continue; diff --git a/source/dnode/vnode/src/tsdb/tsdbCache.c b/source/dnode/vnode/src/tsdb/tsdbCache.c index 291a5ab1eb..d71eb33951 100644 --- a/source/dnode/vnode/src/tsdb/tsdbCache.c +++ b/source/dnode/vnode/src/tsdb/tsdbCache.c @@ -392,7 +392,7 @@ static tb_uid_t getTableSuidByUid(tb_uid_t uid, STsdb *pTsdb) { SMetaReader mr = {0}; metaReaderInit(&mr, pTsdb->pVnode->pMeta, 0); - if (metaGetTableEntryByUid(&mr, uid) < 0) { + if (metaGetTableEntryByUidCache(&mr, uid) < 0) { metaReaderClear(&mr); // table not esist return 0; } diff --git a/source/dnode/vnode/src/tsdb/tsdbRead.c b/source/dnode/vnode/src/tsdb/tsdbRead.c index ca3df04550..96cfa1752d 100644 --- a/source/dnode/vnode/src/tsdb/tsdbRead.c +++ b/source/dnode/vnode/src/tsdb/tsdbRead.c @@ -4350,7 +4350,7 @@ int32_t tsdbGetTableSchema(SVnode* pVnode, int64_t uid, STSchema** pSchema, int6 SMetaReader mr = {0}; metaReaderInit(&mr, pVnode->pMeta, 0); - int32_t code = metaGetTableEntryByUid(&mr, uid); + int32_t code = metaGetTableEntryByUidCache(&mr, uid); if (code != TSDB_CODE_SUCCESS) { terrno = TSDB_CODE_TDB_INVALID_TABLE_ID; metaReaderClear(&mr); @@ -4362,7 +4362,7 @@ int32_t tsdbGetTableSchema(SVnode* pVnode, int64_t uid, STSchema** pSchema, int6 if (mr.me.type == TSDB_CHILD_TABLE) { tDecoderClear(&mr.coder); *suid = mr.me.ctbEntry.suid; - code = metaGetTableEntryByUid(&mr, *suid); + code = metaGetTableEntryByUidCache(&mr, *suid); if (code != TSDB_CODE_SUCCESS) { terrno = TSDB_CODE_TDB_INVALID_TABLE_ID; metaReaderClear(&mr); diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index c072a5f1aa..3219ffa8f7 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -290,7 +290,7 @@ int32_t isQualifiedTable(STableKeyInfo* info, SNode* pTagCond, void* metaHandle, SMetaReader mr = {0}; metaReaderInit(&mr, metaHandle, 0); - code = metaGetTableEntryByUid(&mr, info->uid); + code = metaGetTableEntryByUidCache(&mr, info->uid); if (TSDB_CODE_SUCCESS != code) { metaReaderClear(&mr); *pQualified = false; @@ -1092,7 +1092,7 @@ size_t getTableTagsBufLen(const SNodeList* pGroups) { int32_t getGroupIdFromTagsVal(void* pMeta, uint64_t uid, SNodeList* pGroupNode, char* keyBuf, uint64_t* pGroupId) { SMetaReader mr = {0}; metaReaderInit(&mr, pMeta, 0); - if (metaGetTableEntryByUid(&mr, uid) != 0) { // table not exist + if (metaGetTableEntryByUidCache(&mr, uid) != 0) { // table not exist metaReaderClear(&mr); return TSDB_CODE_PAR_TABLE_NOT_EXIST; } diff --git a/source/libs/executor/src/executor.c b/source/libs/executor/src/executor.c index 1b703c0137..3ee8ea9452 100644 --- a/source/libs/executor/src/executor.c +++ b/source/libs/executor/src/executor.c @@ -287,7 +287,7 @@ static SArray* filterUnqualifiedTables(const SStreamScanInfo* pScanInfo, const S for (int32_t i = 0; i < taosArrayGetSize(tableIdList); ++i) { uint64_t* id = (uint64_t*)taosArrayGet(tableIdList, i); - int32_t code = metaGetTableEntryByUid(&mr, *id); + int32_t code = metaGetTableEntryByUidCache(&mr, *id); if (code != TSDB_CODE_SUCCESS) { qError("failed to get table meta, uid:%" PRIu64 " code:%s, %s", *id, tstrerror(terrno), idstr); continue; diff --git a/source/libs/executor/src/executorimpl.c b/source/libs/executor/src/executorimpl.c index 9ad0aae3e4..db6836b74e 100644 --- a/source/libs/executor/src/executorimpl.c +++ b/source/libs/executor/src/executorimpl.c @@ -2317,7 +2317,7 @@ SSchemaWrapper* extractQueriedColumnSchema(SScanPhysiNode* pScanNode); int32_t extractTableSchemaInfo(SReadHandle* pHandle, SScanPhysiNode* pScanNode, SExecTaskInfo* pTaskInfo) { SMetaReader mr = {0}; metaReaderInit(&mr, pHandle->meta, 0); - int32_t code = metaGetTableEntryByUid(&mr, pScanNode->uid); + int32_t code = metaGetTableEntryByUidCache(&mr, pScanNode->uid); if (code != TSDB_CODE_SUCCESS) { qError("failed to get the table meta, uid:0x%" PRIx64 ", suid:0x%" PRIx64 ", %s", pScanNode->uid, pScanNode->suid, GET_TASKID(pTaskInfo)); @@ -2336,7 +2336,7 @@ int32_t extractTableSchemaInfo(SReadHandle* pHandle, SScanPhysiNode* pScanNode, tDecoderClear(&mr.coder); tb_uid_t suid = mr.me.ctbEntry.suid; - metaGetTableEntryByUid(&mr, suid); + metaGetTableEntryByUidCache(&mr, suid); pSchemaInfo->sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow); pSchemaInfo->tversion = mr.me.stbEntry.schemaTag.version; } else { diff --git a/source/libs/executor/src/scanoperator.c b/source/libs/executor/src/scanoperator.c index 43d09dcac6..84c5bb425b 100644 --- a/source/libs/executor/src/scanoperator.c +++ b/source/libs/executor/src/scanoperator.c @@ -484,7 +484,7 @@ int32_t addTagPseudoColumnData(SReadHandle* pHandle, const SExprInfo* pExpr, int // 1. check if it is existed in meta cache if (pCache == NULL) { metaReaderInit(&mr, pHandle->meta, 0); - code = metaGetTableEntryByUid(&mr, pBlock->info.uid); + code = metaGetTableEntryByUidCache(&mr, pBlock->info.uid); if (code != TSDB_CODE_SUCCESS) { if (terrno == TSDB_CODE_PAR_TABLE_NOT_EXIST) { qWarn("failed to get table meta, table may have been dropped, uid:0x%" PRIx64 ", code:%s, %s", pBlock->info.uid, @@ -508,7 +508,7 @@ int32_t addTagPseudoColumnData(SReadHandle* pHandle, const SExprInfo* pExpr, int h = taosLRUCacheLookup(pCache->pTableMetaEntryCache, &pBlock->info.uid, sizeof(pBlock->info.uid)); if (h == NULL) { metaReaderInit(&mr, pHandle->meta, 0); - code = metaGetTableEntryByUid(&mr, pBlock->info.uid); + code = metaGetTableEntryByUidCache(&mr, pBlock->info.uid); if (code != TSDB_CODE_SUCCESS) { if (terrno == TSDB_CODE_PAR_TABLE_NOT_EXIST) { qWarn("failed to get table meta, table may have been dropped, uid:0x%" PRIx64 ", code:%s, %s", diff --git a/source/libs/executor/src/sysscanoperator.c b/source/libs/executor/src/sysscanoperator.c index c5e1f2c214..eea2549a42 100644 --- a/source/libs/executor/src/sysscanoperator.c +++ b/source/libs/executor/src/sysscanoperator.c @@ -441,6 +441,9 @@ static SSDataBlock* sysTableScanUserTags(SOperatorInfo* pOperator) { int32_t code = metaGetTableEntryByName(&smrChildTable, condTableName); if (code != TSDB_CODE_SUCCESS) { // terrno has been set by metaGetTableEntryByName, therefore, return directly + metaReaderClear(&smrChildTable); + blockDataDestroy(dataBlock); + pInfo->loadInfo.totalRows = 0; return NULL; } @@ -456,12 +459,16 @@ static SSDataBlock* sysTableScanUserTags(SOperatorInfo* pOperator) { code = metaGetTableEntryByUid(&smrSuperTable, smrChildTable.me.ctbEntry.suid); if (code != TSDB_CODE_SUCCESS) { // terrno has been set by metaGetTableEntryByUid + metaReaderClear(&smrSuperTable); + metaReaderClear(&smrChildTable); + blockDataDestroy(dataBlock); return NULL; } sysTableUserTagsFillOneTableTags(pInfo, &smrSuperTable, &smrChildTable, dbname, tableName, &numOfRows, dataBlock); metaReaderClear(&smrSuperTable); metaReaderClear(&smrChildTable); + if (numOfRows > 0) { relocateAndFilterSysTagsScanResult(pInfo, numOfRows, dataBlock, pOperator->exprSupp.pFilterInfo); numOfRows = 0; From 31c2aa89efd5443f60db0061995bd2abbbaa9db3 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sun, 27 Nov 2022 01:59:49 +0800 Subject: [PATCH 63/83] fix(query): do some internal refactor. --- source/libs/executor/src/executil.c | 168 +++++++++++------- .../libs/function/src/detail/tavgfunction.c | 26 ++- tests/script/tsim/testsuit.sim | 48 ++--- 3 files changed, 136 insertions(+), 106 deletions(-) diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 3219ffa8f7..65937ae1bc 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -405,7 +405,7 @@ static SColumnInfoData* getColInfoResult(void* metaHandle, int64_t suid, SArray* terrno = TSDB_CODE_OUT_OF_MEMORY; goto end; } - ctx.index = 0; + ctx.cInfoList = taosArrayInit(4, sizeof(SColumnInfo)); if (ctx.cInfoList == NULL) { terrno = TSDB_CODE_OUT_OF_MEMORY; @@ -964,34 +964,98 @@ static int32_t optimizeTbnameInCondImpl(void* metaHandle, int64_t suid, SArray* return -1; } +static void genTagFilterDigest(const SNode* pTagCond, T_MD5_CTX* pContext) { + if (pTagCond == NULL) { + return; + } + + char* payload = NULL; + int32_t len = 0; + nodesNodeToMsg(pTagCond, &payload, &len); + + tMD5Init(pContext); + tMD5Update(pContext, (uint8_t*)payload, (uint32_t)len); + tMD5Final(pContext); + + taosMemoryFree(payload); +} + +static int32_t doFilterByTagCond(STableListInfo* pListInfo, SArray* res, SNode* pTagCond, void* metaHandle) { + if (pTagCond == NULL) { + return TSDB_CODE_SUCCESS; + } + + terrno = TDB_CODE_SUCCESS; + SColumnInfoData* pColInfoData = getColInfoResult(metaHandle, pListInfo->suid, res, pTagCond); + if (terrno != TDB_CODE_SUCCESS) { + colDataDestroy(pColInfoData); + taosMemoryFreeClear(pColInfoData); + taosArrayDestroy(res); + qError("failed to getColInfoResult, code: %s", tstrerror(terrno)); + return terrno; + } + + int32_t i = 0; + int32_t len = taosArrayGetSize(res); + + if (pColInfoData != NULL) { + bool* pResult = (bool*)pColInfoData->pData; + SArray* p = taosArrayInit(taosArrayGetSize(res), sizeof(uint64_t)); + + while (i < len && pColInfoData) { + int64_t* uid = taosArrayGet(res, i); + qDebug("tagfilter get uid:%" PRId64 ", res:%d", *uid, pResult[i]); + + if (pResult[i]) { + taosArrayPush(p, uid); + } + i += 1; + } + + taosArraySwap(res, p); + taosArrayDestroy(p); + } + + colDataDestroy(pColInfoData); + taosMemoryFreeClear(pColInfoData); + + return TSDB_CODE_SUCCESS; +} + int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, SNode* pTagCond, SNode* pTagIndexCond, STableListInfo* pListInfo) { int32_t code = TSDB_CODE_SUCCESS; + size_t numOfTables = 0; uint64_t tableUid = pScanNode->uid; pListInfo->suid = pScanNode->suid; SArray* res = taosArrayInit(8, sizeof(uint64_t)); - if (pScanNode->tableType == TSDB_SUPER_TABLE) { - // try to retrieve the result from meta cache - // generate the cache key - T_MD5_CTX context = {0}; - - if (pTagIndexCond) { - char* payload = NULL; - int32_t len = 0; - nodesNodeToMsg(pTagCond, &payload, &len); - - tMD5Init(&context); - tMD5Update(&context, (uint8_t*)payload, (uint32_t)len); - tMD5Final(&context); - - taosMemoryFree(payload); + if (pScanNode->tableType != TSDB_SUPER_TABLE) { + if (metaIsTableExist(metaHandle, tableUid)) { + taosArrayPush(res, &tableUid); } + code = doFilterByTagCond(pListInfo, res, pTagCond, metaHandle); + if (code != TSDB_CODE_SUCCESS) { + return code; + } + } else { + // try to retrieve the result from meta cache + T_MD5_CTX context = {0}; + genTagFilterDigest(pTagCond, &context); + bool acquired = false; metaGetCachedTableUidList(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), res, &acquired); - if (!acquired) { + if (acquired) { + qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t)taosArrayGetSize(res)); + goto _end; + } + + if (!pTagCond) { // no tag condition exists, let's fetch all tables of this super table + ASSERT(pTagIndexCond == NULL); + vnodeGetCtbIdList(pVnode, pScanNode->suid, res); + } else { // failed to find the result in the cache, let try to calculate the results if (pTagIndexCond) { SIndexMetaArg metaArg = { @@ -1003,63 +1067,29 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, qError("failed to get tableIds from index, reason:%s, suid:%" PRIu64, tstrerror(code), tableUid); code = TDB_CODE_SUCCESS; } - } else if (!pTagCond) { - vnodeGetCtbIdList(pVnode, pScanNode->suid, res); } - - // let's add the filter results into meta-cache - size_t numOfTables = taosArrayGetSize(res); - size_t size = numOfTables * sizeof(uint64_t) + sizeof(int32_t); - char* pPayload = taosMemoryMalloc(size); - *(int32_t*)pPayload = numOfTables; - - if (numOfTables > 0) { - memcpy(pPayload + sizeof(int32_t), taosArrayGet(res, 0), numOfTables * sizeof(uint64_t)); - } - - metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, - size, 1); - } else { - qDebug("retrieve table uid list from cache, numOfTables:%d", (int32_t) taosArrayGetSize(res)); } - } else { // Create one table group. - if (metaIsTableExist(metaHandle, tableUid)) { - taosArrayPush(res, &tableUid); + + code = doFilterByTagCond(pListInfo, res, pTagCond, metaHandle); + if (code != TSDB_CODE_SUCCESS) { + return code; } + + // let's add the filter results into meta-cache + numOfTables = taosArrayGetSize(res); + size_t size = numOfTables * sizeof(uint64_t) + sizeof(int32_t); + char* pPayload = taosMemoryMalloc(size); + *(int32_t*)pPayload = numOfTables; + + if (numOfTables > 0) { + memcpy(pPayload + sizeof(int32_t), taosArrayGet(res, 0), numOfTables * sizeof(uint64_t)); + } + + metaUidFilterCachePut(metaHandle, pScanNode->suid, context.digest, tListLen(context.digest), pPayload, size, 1); } - if (pTagCond) { - terrno = TDB_CODE_SUCCESS; - SColumnInfoData* pColInfoData = getColInfoResult(metaHandle, pListInfo->suid, res, pTagCond); - if (terrno != TDB_CODE_SUCCESS) { - colDataDestroy(pColInfoData); - taosMemoryFreeClear(pColInfoData); - taosArrayDestroy(res); - qError("failed to getColInfoResult, code: %s", tstrerror(terrno)); - return terrno; - } - - int32_t i = 0; - int32_t j = 0; - int32_t len = taosArrayGetSize(res); - while (i < taosArrayGetSize(res) && j < len && pColInfoData) { - void* var = POINTER_SHIFT(pColInfoData->pData, j * pColInfoData->info.bytes); - - int64_t* uid = taosArrayGet(res, i); - qDebug("tagfilter get uid:%" PRId64 ", res:%d", *uid, *(bool*)var); - if (*(bool*)var == false) { - taosArrayRemove(res, i); - j++; - continue; - } - i++; - j++; - } - colDataDestroy(pColInfoData); - taosMemoryFreeClear(pColInfoData); - } - - size_t numOfTables = taosArrayGetSize(res); +_end: + numOfTables = taosArrayGetSize(res); for (int i = 0; i < numOfTables; i++) { STableKeyInfo info = {.uid = *(uint64_t*)taosArrayGet(res, i), .groupId = 0}; diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 9507380d38..17190d8436 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -708,24 +708,22 @@ int32_t avgCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx) { } int32_t avgFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { - SInputColumnInfoData* pInput = &pCtx->input; + SResultRowEntryInfo* pEntryInfo = GET_RES_INFO(pCtx); - SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); - int32_t type = pAvgRes->type; + SAvgRes* pRes = GET_ROWCELL_INTERBUF(pEntryInfo); + int32_t type = pRes->type; - if (IS_SIGNED_NUMERIC_TYPE(type)) { - pAvgRes->result = pAvgRes->sum.isum / ((double)pAvgRes->count); - } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { - pAvgRes->result = pAvgRes->sum.usum / ((double)pAvgRes->count); - } else { - pAvgRes->result = pAvgRes->sum.dsum / ((double)pAvgRes->count); - } - - // check for overflow - if (isinf(pAvgRes->result) || isnan(pAvgRes->result)) { - GET_RES_INFO(pCtx)->numOfRes = 0; + if (pRes->count > 0) { + if (IS_SIGNED_NUMERIC_TYPE(type)) { + pRes->result = pRes->sum.isum / ((double)pRes->count); + } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { + pRes->result = pRes->sum.usum / ((double)pRes->count); + } else { + pRes->result = pRes->sum.dsum / ((double)pRes->count); + } } + pEntryInfo->numOfRes = (pRes->count > 0)? 1:0; return functionFinalize(pCtx, pBlock); } diff --git a/tests/script/tsim/testsuit.sim b/tests/script/tsim/testsuit.sim index ad8d70b089..c5fbf41b66 100644 --- a/tests/script/tsim/testsuit.sim +++ b/tests/script/tsim/testsuit.sim @@ -5,30 +5,32 @@ #run tsim/table/basic1.sim #run tsim/trans/lossdata1.sim #run tsim/trans/create_db.sim -run tsim/stable/alter_metrics.sim -run tsim/stable/tag_modify.sim -run tsim/stable/alter_comment.sim -run tsim/stable/column_drop.sim -run tsim/stable/column_modify.sim -run tsim/stable/tag_rename.sim -run tsim/stable/vnode3.sim -run tsim/stable/metrics.sim -run tsim/stable/alter_insert2.sim -run tsim/stable/show.sim -run tsim/stable/alter_import.sim -run tsim/stable/tag_add.sim -run tsim/stable/tag_drop.sim -run tsim/stable/column_add.sim -run tsim/stable/alter_count.sim -run tsim/stable/values.sim -run tsim/stable/dnode3.sim -run tsim/stable/alter_insert1.sim -run tsim/stable/refcount.sim -run tsim/stable/tag_filter.sim -run tsim/stable/disk.sim -run tsim/db/basic1.sim + +#run tsim/stable/alter_metrics.sim +#run tsim/stable/tag_modify.sim +#run tsim/stable/alter_comment.sim +#run tsim/stable/column_drop.sim +#run tsim/stable/column_modify.sim +#run tsim/stable/tag_rename.sim +#run tsim/stable/vnode3.sim +#run tsim/stable/metrics.sim +#run tsim/stable/alter_insert2.sim +#run tsim/stable/alter_import.sim +#run tsim/stable/tag_add.sim +#run tsim/stable/tag_drop.sim +#run tsim/stable/column_add.sim +#run tsim/stable/alter_count.sim +#run tsim/stable/values.sim +#run tsim/stable/dnode3.sim +#run tsim/stable/alter_insert1.sim +#run tsim/stable/refcount.sim +#run tsim/stable/tag_filter.sim +#run tsim/stable/disk.sim +#run tsim/db/basic1.sim +run tsim/db/basic2.sim run tsim/db/basic3.sim -run tsim/db/basic7.sim +run tsim/db/basic4.sim +run tsim/db/basic5.sim run tsim/db/basic6.sim run tsim/db/alter_replica_13.sim run tsim/db/create_all_options.sim From 203ea1f9c8c665d5732ba985b344cda6833c70e5 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sun, 27 Nov 2022 14:57:44 +0800 Subject: [PATCH 64/83] fix(query): fix error --- include/libs/executor/executor.h | 3 +++ source/common/src/tdatablock.c | 4 ---- source/dnode/vnode/src/sma/smaRollup.c | 6 ++++-- source/libs/executor/src/exchangeoperator.c | 1 + source/libs/executor/src/executor.c | 12 ++++++++++++ source/libs/executor/src/groupoperator.c | 5 ++--- source/libs/function/src/detail/tavgfunction.c | 2 +- source/libs/function/src/detail/tminmax.c | 2 +- tests/system-test/1-insert/block_wise.py | 2 +- tests/system-test/2-query/avg.py | 6 ++++-- tests/system-test/2-query/max_partition.py | 2 +- 11 files changed, 30 insertions(+), 15 deletions(-) diff --git a/include/libs/executor/executor.h b/include/libs/executor/executor.h index 1fe30a2d66..0bca254e14 100644 --- a/include/libs/executor/executor.h +++ b/include/libs/executor/executor.h @@ -142,8 +142,11 @@ int32_t qGetQueryTableSchemaVersion(qTaskInfo_t tinfo, char* dbName, char* table */ int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, bool* hasMore, SLocalFetch* pLocal); + int32_t qExecTask(qTaskInfo_t tinfo, SSDataBlock** pBlock, uint64_t* useconds); +void qCleanExecTaskBlockBuf(qTaskInfo_t tinfo); + /** * kill the ongoing query asynchronously * @param tinfo qhandle diff --git a/source/common/src/tdatablock.c b/source/common/src/tdatablock.c index 6a71d441b0..cfa2964e16 100644 --- a/source/common/src/tdatablock.c +++ b/source/common/src/tdatablock.c @@ -1322,10 +1322,6 @@ int32_t copyDataBlock(SSDataBlock* dst, const SSDataBlock* src) { for (int32_t i = 0; i < numOfCols; ++i) { SColumnInfoData* pDst = taosArrayGet(dst->pDataBlock, i); SColumnInfoData* pSrc = taosArrayGet(src->pDataBlock, i); - if (pSrc->pData == NULL) { - continue; - } - colDataAssign(pDst, pSrc, src->info.rows, &src->info); } diff --git a/source/dnode/vnode/src/sma/smaRollup.c b/source/dnode/vnode/src/sma/smaRollup.c index c5f040c987..51842a8ae4 100644 --- a/source/dnode/vnode/src/sma/smaRollup.c +++ b/source/dnode/vnode/src/sma/smaRollup.c @@ -739,11 +739,13 @@ static int32_t tdRSmaExecAndSubmitResult(SSma *pSma, qTaskInfo_t taskInfo, SRSma } } - tdBlockDataDestroy(pResList); + taosArrayDestroy(pResList); + qCleanExecTaskBlockBuf(taskInfo); return TSDB_CODE_SUCCESS; _err: - tdBlockDataDestroy(pResList); + taosArrayDestroy(pResList); + qCleanExecTaskBlockBuf(taskInfo); return TSDB_CODE_FAILED; } diff --git a/source/libs/executor/src/exchangeoperator.c b/source/libs/executor/src/exchangeoperator.c index de918e0034..d8c85c5ffb 100644 --- a/source/libs/executor/src/exchangeoperator.c +++ b/source/libs/executor/src/exchangeoperator.c @@ -646,6 +646,7 @@ int32_t seqLoadRemoteData(SOperatorInfo* pOperator) { SRetrieveTableRsp* pRsp = pDataInfo->pRsp; SLoadRemoteDataInfo* pLoadInfo = &pExchangeInfo->loadInfo; + if (pRsp->numOfRows == 0) { qDebug("%s vgId:%d, taskID:0x%" PRIx64 " execId:%d %d of total completed, rowsOfSource:%" PRIu64 ", totalRows:%" PRIu64 " try next", diff --git a/source/libs/executor/src/executor.c b/source/libs/executor/src/executor.c index 3ee8ea9452..01baba5a52 100644 --- a/source/libs/executor/src/executor.c +++ b/source/libs/executor/src/executor.c @@ -572,6 +572,18 @@ int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, bo return pTaskInfo->code; } +void qCleanExecTaskBlockBuf(qTaskInfo_t tinfo) { + SExecTaskInfo* pTaskInfo = (SExecTaskInfo*)tinfo; + SArray* pList = pTaskInfo->pResultBlockList; + size_t num = taosArrayGetSize(pList); + for(int32_t i = 0; i < num; ++i) { + SSDataBlock** p = taosArrayGet(pTaskInfo->pResultBlockList, i); + blockDataDestroy(*p); + } + + taosArrayClear(pTaskInfo->pResultBlockList); +} + int32_t qExecTask(qTaskInfo_t tinfo, SSDataBlock** pRes, uint64_t* useconds) { SExecTaskInfo* pTaskInfo = (SExecTaskInfo*)tinfo; int64_t threadId = taosGetSelfPthreadId(); diff --git a/source/libs/executor/src/groupoperator.c b/source/libs/executor/src/groupoperator.c index cde8346487..066912fbdf 100644 --- a/source/libs/executor/src/groupoperator.c +++ b/source/libs/executor/src/groupoperator.c @@ -274,10 +274,9 @@ static void doHashGroupbyAgg(SOperatorInfo* pOperator, SSDataBlock* pBlock) { // return; // } - int32_t len = 0; - STimeWindow w = TSWINDOW_INITIALIZER; - + int32_t len = 0; terrno = TSDB_CODE_SUCCESS; + int32_t num = 0; for (int32_t j = 0; j < pBlock->info.rows; ++j) { // Compare with the previous row of this column, and do not set the output buffer again if they are identical. diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 17190d8436..267cb36769 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -494,7 +494,7 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) { switch(type) { case TSDB_DATA_TYPE_UTINYINT: case TSDB_DATA_TYPE_TINYINT: { - const int8_t* plist = (const int8_t*) &pCol->pData[start]; + const int8_t* plist = (const int8_t*) pCol->pData; // 1. If the CPU supports AVX, let's employ AVX instructions to speedup this loop if (simdAvailable) { diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index 46e68f46ec..e47edb8a1e 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -827,7 +827,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { if (i >= end) { ASSERT(numOfElems == 0); - return numOfElems; + goto _over; } doExtractVal(pCol, i, end, pCtx, pBuf, isMinFunc); diff --git a/tests/system-test/1-insert/block_wise.py b/tests/system-test/1-insert/block_wise.py index 0e17a01d05..8222000cd6 100644 --- a/tests/system-test/1-insert/block_wise.py +++ b/tests/system-test/1-insert/block_wise.py @@ -145,7 +145,7 @@ class TDTestCase: def init(self, conn, logSql, replicaVar=1): self.replicaVar = int(replicaVar) tdLog.debug(f"start to excute {__file__}") - tdSql.init(conn.cursor(), False) + tdSql.init(conn.cursor(), True) self.precision = "ms" self.sma_count = 0 self.sma_created_index = [] diff --git a/tests/system-test/2-query/avg.py b/tests/system-test/2-query/avg.py index 910dd524cb..139e7d4bf4 100644 --- a/tests/system-test/2-query/avg.py +++ b/tests/system-test/2-query/avg.py @@ -1,5 +1,7 @@ import taos import sys + +import math import numpy as np from util.log import * from util.sql import * @@ -411,7 +413,7 @@ class TDTestCase: tdSql.checkData(0,2,14042.142857143) tdSql.checkData(0,3,53.571428571) tdSql.checkData(0,4,5.828571332045761e+37) - tdSql.checkData(0,5,None) + tdSql.checkData(0,5,math.inf) # check + - * / in functions @@ -421,7 +423,7 @@ class TDTestCase: tdSql.checkData(0,2,14042.142857143) tdSql.checkData(0,3,26.785714286) tdSql.checkData(0,4,2.9142856660228804e+37) - tdSql.checkData(0,5,None) + tdSql.checkData(0,5,math.inf) diff --git a/tests/system-test/2-query/max_partition.py b/tests/system-test/2-query/max_partition.py index b14bc649dd..dec24010fc 100644 --- a/tests/system-test/2-query/max_partition.py +++ b/tests/system-test/2-query/max_partition.py @@ -7,7 +7,7 @@ class TDTestCase: def init(self, conn, logSql, replicaVar=1): self.replicaVar = int(replicaVar) tdLog.debug("start to execute %s" % __file__) - tdSql.init(conn.cursor()) + tdSql.init(conn.cursor(), True) self.row_nums = 10 self.tb_nums = 10 From c207caf94994267fe46b1108568ce258253c09bc Mon Sep 17 00:00:00 2001 From: Sai Vishwak <55258956+saivishwak@users.noreply.github.com> Date: Sun, 27 Nov 2022 13:00:53 +0530 Subject: [PATCH 65/83] Update number of stars in README --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 8d2567a816..6aec756ec7 100644 --- a/README.md +++ b/README.md @@ -31,7 +31,7 @@ TDengine is an open source, high-performance, cloud native [time-series database - **[Easy Data Analytics](https://tdengine.com/tdengine/time-series-data-analytics-made-easy/)**: Through super tables, storage and compute separation, data partitioning by time interval, pre-computation and other means, TDengine makes it easy to explore, format, and get access to data in a highly efficient way. -- **[Open Source](https://tdengine.com/tdengine/open-source-time-series-database/)**: TDengine’s core modules, including cluster feature, are all available under open source licenses. It has gathered 18.8k stars on GitHub. There is an active developer community, and over 139k running instances worldwide. +- **[Open Source](https://tdengine.com/tdengine/open-source-time-series-database/)**: TDengine’s core modules, including cluster feature, are all available under open source licenses. It has gathered 19.9k stars on GitHub. There is an active developer community, and over 139k running instances worldwide. For a full list of TDengine competitive advantages, please [check here](https://tdengine.com/tdengine/). The easiest way to experience TDengine is through [TDengine Cloud](https://cloud.tdengine.com). From f08a183ab597d2108aa31e1755f0395726381d1c Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 15:39:01 +0800 Subject: [PATCH 66/83] test: fsync consum file --- utils/test/c/tmqSim.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/utils/test/c/tmqSim.c b/utils/test/c/tmqSim.c index f77a910f35..fe2bb59af3 100644 --- a/utils/test/c/tmqSim.c +++ b/utils/test/c/tmqSim.c @@ -865,6 +865,9 @@ void loop_consume(SThreadInfo* pInfo) { taosFprintfFile(g_fp, "==== consumerId: %d, consumeMsgCnt: %" PRId64 ", consumeRowCnt: %" PRId64 "\n", pInfo->consumerId, pInfo->consumeMsgCnt, pInfo->consumeRowCnt); + + taosFsyncFile(pInfo->pConsumeRowsFile); + taosCloseFile(&pInfo->pConsumeRowsFile); } void* consumeThreadFunc(void* param) { From 4cd660b3b12d8ad60cbcd016584fa52dd7bf277d Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 15:39:17 +0800 Subject: [PATCH 67/83] refactor: format tmqSim.c --- utils/test/c/tmqSim.c | 511 ++++++++++++++++++++---------------------- 1 file changed, 247 insertions(+), 264 deletions(-) diff --git a/utils/test/c/tmqSim.c b/utils/test/c/tmqSim.c index fe2bb59af3..9c1dc2e063 100644 --- a/utils/test/c/tmqSim.c +++ b/utils/test/c/tmqSim.c @@ -14,13 +14,13 @@ */ #include +#include #include #include #include #include #include #include -#include #include "taos.h" #include "taosdef.h" @@ -36,7 +36,7 @@ #define MAX_ROW_STR_LEN (16 * 1024) #define MAX_CONSUMER_THREAD_CNT (16) #define MAX_VGROUP_CNT (32) -#define SEND_TIME_UNIT 10 // ms +#define SEND_TIME_UNIT 10 // ms #define MAX_SQL_LEN 1048576 typedef enum { @@ -45,11 +45,7 @@ typedef enum { NOTIFY_CMD_ID_BUTT, } NOTIFY_CMD_ID; -typedef enum enumQUERY_TYPE { - NO_INSERT_TYPE, - INSERT_TYPE, - QUERY_TYPE_BUT -} QUERY_TYPE; +typedef enum enumQUERY_TYPE { NO_INSERT_TYPE, INSERT_TYPE, QUERY_TYPE_BUT } QUERY_TYPE; typedef struct { TdThread thread; @@ -61,7 +57,7 @@ typedef struct { // char autoOffsetRest[16]; // none, earliest, latest TdFilePtr pConsumeRowsFile; - TdFilePtr pConsumeMetaFile; + TdFilePtr pConsumeMetaFile; int32_t ifCheckData; int64_t expectMsgCnt; @@ -87,12 +83,12 @@ typedef struct { int32_t rowsOfPerVgroups[MAX_VGROUP_CNT][2]; // [i][0]: vgroup id, [i][1]: rows of consume int64_t ts; - TAOS* taos; + TAOS* taos; // below parameters is used by omb test - int32_t producerRate; // unit: msgs/s - int64_t totalProduceMsgs; - int64_t totalMsgsLen; + int32_t producerRate; // unit: msgs/s + int64_t totalProduceMsgs; + int64_t totalMsgsLen; } SThreadInfo; @@ -112,12 +108,12 @@ typedef struct { SThreadInfo stProdThreads[MAX_CONSUMER_THREAD_CNT]; // below parameters is used by omb test - char topic[64]; - int32_t producers; - int32_t producerRate; - int32_t runDurationMinutes; - int32_t batchSize; - int32_t payloadLen; + char topic[64]; + int32_t producers; + int32_t producerRate; + int32_t runDurationMinutes; + int32_t batchSize; + int32_t payloadLen; } SConfInfo; static SConfInfo g_stConfInfo; @@ -146,14 +142,13 @@ static void printHelp() { printf("%s%s%s%ds\n", indent, indent, "consume delay, default is ", g_stConfInfo.consumeDelay); printf("%s%s\n", indent, "-e"); printf("%s%s%s%d\n", indent, indent, "snapshot, default is ", g_stConfInfo.useSnapshot); - + printf("%s%s\n", indent, "-t"); printf("%s%s%s\n", indent, indent, "topic name, default is null"); printf("%s%s\n", indent, "-x"); printf("%s%s%s\n", indent, indent, "consume thread number, default is 1"); - printf("%s%s\n", indent, "-l"); printf("%s%s%s%d\n", indent, indent, "run duration unit is minutes, default is ", g_stConfInfo.runDurationMinutes); printf("%s%s\n", indent, "-p"); @@ -165,7 +160,6 @@ static void printHelp() { printf("%s%s\n", indent, "-n"); printf("%s%s%s\n", indent, indent, "payload len unit is byte, default is 1000"); - exit(EXIT_SUCCESS); } @@ -194,7 +188,7 @@ void initLogFile() { pid_t process_id = getpid(); if (0 != strlen(g_stConfInfo.topic)) { - sprintf(filename, "/tmp/tmqlog-%d-%s.txt", process_id, getCurrentTimeString(tmpString)); + sprintf(filename, "/tmp/tmqlog-%d-%s.txt", process_id, getCurrentTimeString(tmpString)); } else { sprintf(filename, "%s/../log/tmqlog-%d-%s.txt", configDir, process_id, getCurrentTimeString(tmpString)); } @@ -294,7 +288,7 @@ void parseArgument(int32_t argc, char* argv[]) { g_stConfInfo.producerRate = atol(argv[++i]); } else if (strcmp(argv[i], "-n") == 0) { g_stConfInfo.payloadLen = atol(argv[++i]); - if(g_stConfInfo.payloadLen <= 0 || g_stConfInfo.payloadLen > 1024 * 1024 * 1024){ + if (g_stConfInfo.payloadLen <= 0 || g_stConfInfo.payloadLen > 1024 * 1024 * 1024) { pError("%s calloc size is too large: %s %s", GREEN, argv[++i], NC); exit(-1); } @@ -357,9 +351,9 @@ void ltrim(char* str) { } int queryDB(TAOS* taos, char* command) { - int retryCnt = 10; - int code = 0; - TAOS_RES* pRes = NULL; + int retryCnt = 10; + int code = 0; + TAOS_RES* pRes = NULL; while (retryCnt--) { pRes = taos_query(taos, command); @@ -379,7 +373,6 @@ int queryDB(TAOS* taos, char* command) { return -1; } - void addRowsToVgroupId(SThreadInfo* pInfo, int32_t vgroupId, int32_t rows) { int32_t i; for (i = 0; i < pInfo->numOfVgroups; i++) { @@ -403,22 +396,21 @@ void addRowsToVgroupId(SThreadInfo* pInfo, int32_t vgroupId, int32_t rows) { } TAOS* createNewTaosConnect() { - TAOS* taos = NULL; - int32_t retryCnt = 10; + TAOS* taos = NULL; + int32_t retryCnt = 10; while (retryCnt--) { TAOS* taos = taos_connect(NULL, "root", "taosdata", NULL, 0); - if (NULL != taos) { - return taos; - } - taosSsleep(1); + if (NULL != taos) { + return taos; + } + taosSsleep(1); } taosFprintfFile(g_fp, "taos_connect() fail\n"); return NULL; } - int32_t saveConsumeContentToTbl(SThreadInfo* pInfo, char* buf) { char sqlStr[1100] = {0}; @@ -440,7 +432,7 @@ int32_t saveConsumeContentToTbl(SThreadInfo* pInfo, char* buf) { if (retCode != 0) { taosFprintfFile(g_fp, "error in save consume content\n"); taosCloseFile(&g_fp); - taos_close(pConn); + taos_close(pConn); exit(-1); } @@ -481,7 +473,7 @@ static char* shellFormatTimestamp(char* buf, int64_t val, int32_t precision) { struct tm ptm; taosLocalTime(&tt, &ptm); - size_t pos = strftime(buf, 35, "%Y-%m-%d %H:%M:%S", &ptm); + size_t pos = strftime(buf, 35, "%Y-%m-%d %H:%M:%S", &ptm); if (precision == TSDB_TIME_PRECISION_NANO) { sprintf(buf + pos, ".%09d", ms); @@ -548,22 +540,20 @@ static void shellDumpFieldToFile(TdFilePtr pFile, const char* val, TAOS_FIELD* f break; case TSDB_DATA_TYPE_BINARY: case TSDB_DATA_TYPE_NCHAR: - case TSDB_DATA_TYPE_JSON: - { - int32_t bufIndex = 0; - for (int32_t i = 0; i < length; i++) { + case TSDB_DATA_TYPE_JSON: { + int32_t bufIndex = 0; + for (int32_t i = 0; i < length; i++) { + buf[bufIndex] = val[i]; + bufIndex++; + if (val[i] == '\"') { buf[bufIndex] = val[i]; bufIndex++; - if (val[i] == '\"') { - buf[bufIndex] = val[i]; - bufIndex++; - } } - buf[bufIndex] = 0; - - taosFprintfFile(pFile, "%s%s%s", quotationStr, buf, quotationStr); } - break; + buf[bufIndex] = 0; + + taosFprintfFile(pFile, "%s%s%s", quotationStr, buf, quotationStr); + } break; case TSDB_DATA_TYPE_TIMESTAMP: shellFormatTimestamp(buf, *(int64_t*)val, precision); taosFprintfFile(pFile, "%s%s%s", quotationStr, buf, quotationStr); @@ -635,7 +625,6 @@ static int32_t data_msg_process(TAOS_RES* msg, SThreadInfo* pInfo, int32_t msgIn return totalRows; } - static int32_t meta_msg_process(TAOS_RES* msg, SThreadInfo* pInfo, int32_t msgIndex) { char buf[1024]; int32_t totalRows = 0; @@ -650,24 +639,24 @@ static int32_t meta_msg_process(TAOS_RES* msg, SThreadInfo* pInfo, int32_t msgIn { tmq_raw_data raw = {0}; - int32_t code = tmq_get_raw(msg, &raw); - - if(code == TSDB_CODE_SUCCESS){ -// int retCode = queryDB(pInfo->taos, "use metadb"); -// if (retCode != 0) { -// taosFprintfFile(g_fp, "error when use metadb\n"); -// taosCloseFile(&g_fp); -// exit(-1); -// } -// taosFprintfFile(g_fp, "raw:%p\n", &raw); -// -// tmq_write_raw(pInfo->taos, raw); + int32_t code = tmq_get_raw(msg, &raw); + + if (code == TSDB_CODE_SUCCESS) { + // int retCode = queryDB(pInfo->taos, "use metadb"); + // if (retCode != 0) { + // taosFprintfFile(g_fp, "error when use metadb\n"); + // taosCloseFile(&g_fp); + // exit(-1); + // } + // taosFprintfFile(g_fp, "raw:%p\n", &raw); + // + // tmq_write_raw(pInfo->taos, raw); } - + char* result = tmq_get_json_meta(msg); - if(result && strcmp(result, "") != 0){ - //printf("meta result: %s\n", result); - taosFprintfFile(pInfo->pConsumeMetaFile, "%s\n", result); + if (result && strcmp(result, "") != 0) { + // printf("meta result: %s\n", result); + taosFprintfFile(pInfo->pConsumeMetaFile, "%s\n", result); } tmq_free_json_meta(result); } @@ -683,8 +672,8 @@ int32_t notifyMainScript(SThreadInfo* pInfo, int32_t cmdId) { char sqlStr[1024] = {0}; // schema: ts timestamp, consumerid int, consummsgcnt bigint, checkresult int - sprintf(sqlStr, "insert into %s.notifyinfo values (%" PRId64 ", %d, %d)", g_stConfInfo.cdbName, atomic_fetch_add_64(&g_stConfInfo.nowTime, 1), cmdId, - pInfo->consumerId); + sprintf(sqlStr, "insert into %s.notifyinfo values (%" PRId64 ", %d, %d)", g_stConfInfo.cdbName, + atomic_fetch_add_64(&g_stConfInfo.nowTime, 1), cmdId, pInfo->consumerId); taos_query_a(pInfo->taos, sqlStr, appNothing, NULL); @@ -695,15 +684,15 @@ int32_t notifyMainScript(SThreadInfo* pInfo, int32_t cmdId) { static int32_t g_once_commit_flag = 0; static void tmq_commit_cb_print(tmq_t* tmq, int32_t code, void* param) { - taosFprintfFile(g_fp, "tmq_commit_cb_print() commit %d\n", code); + taosFprintfFile(g_fp, "tmq_commit_cb_print() commit %d\n", code); - if (0 == g_once_commit_flag) { - g_once_commit_flag = 1; - notifyMainScript((SThreadInfo*)param, (int32_t)NOTIFY_CMD_START_COMMIT); + if (0 == g_once_commit_flag) { + g_once_commit_flag = 1; + notifyMainScript((SThreadInfo*)param, (int32_t)NOTIFY_CMD_START_COMMIT); } - char tmpString[128]; - taosFprintfFile(g_fp, "%s tmq_commit_cb_print() be called\n", getCurrentTimeString(tmpString)); + char tmpString[128]; + taosFprintfFile(g_fp, "%s tmq_commit_cb_print() be called\n", getCurrentTimeString(tmpString)); } void build_consumer(SThreadInfo* pInfo) { @@ -768,7 +757,7 @@ int32_t saveConsumeResult(SThreadInfo* pInfo) { int retCode = queryDB(pInfo->taos, sqlStr); if (retCode != 0) { - taosFprintfFile(g_fp, "consume id %d error in save consume result\n", pInfo->consumerId); + taosFprintfFile(g_fp, "consume id %d error in save consume result\n", pInfo->consumerId); return -1; } @@ -797,9 +786,9 @@ void loop_consume(SThreadInfo* pInfo) { sprintf(filename, "%s/../log/consumerid_%d.txt", configDir, pInfo->consumerId); pInfo->pConsumeRowsFile = taosOpenFile(filename, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC | TD_FILE_STREAM); - sprintf(filename, "%s/../log/meta_consumerid_%d.txt", configDir, pInfo->consumerId); - pInfo->pConsumeMetaFile = taosOpenFile(filename, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC | TD_FILE_STREAM); - + sprintf(filename, "%s/../log/meta_consumerid_%d.txt", configDir, pInfo->consumerId); + pInfo->pConsumeMetaFile = taosOpenFile(filename, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC | TD_FILE_STREAM); + if (pInfo->pConsumeRowsFile == NULL || pInfo->pConsumeMetaFile == NULL) { taosFprintfFile(g_fp, "%s create file fail for save rows or save meta\n", getCurrentTimeString(tmpString)); return; @@ -815,15 +804,15 @@ void loop_consume(SThreadInfo* pInfo) { TAOS_RES* tmqMsg = tmq_consumer_poll(pInfo->tmq, consumeDelay); if (tmqMsg) { if (0 != g_stConfInfo.showMsgFlag) { - tmq_res_t msgType = tmq_get_res_type(tmqMsg); - if (msgType == TMQ_RES_TABLE_META) { - totalRows += meta_msg_process(tmqMsg, pInfo, totalMsgs); - } else if (msgType == TMQ_RES_DATA){ - totalRows += data_msg_process(tmqMsg, pInfo, totalMsgs); - } else if (msgType == TMQ_RES_METADATA){ - meta_msg_process(tmqMsg, pInfo, totalMsgs); - totalRows += data_msg_process(tmqMsg, pInfo, totalMsgs); - } + tmq_res_t msgType = tmq_get_res_type(tmqMsg); + if (msgType == TMQ_RES_TABLE_META) { + totalRows += meta_msg_process(tmqMsg, pInfo, totalMsgs); + } else if (msgType == TMQ_RES_DATA) { + totalRows += data_msg_process(tmqMsg, pInfo, totalMsgs); + } else if (msgType == TMQ_RES_METADATA) { + meta_msg_process(tmqMsg, pInfo, totalMsgs); + totalRows += data_msg_process(tmqMsg, pInfo, totalMsgs); + } } taos_free_result(tmqMsg); @@ -865,7 +854,7 @@ void loop_consume(SThreadInfo* pInfo) { taosFprintfFile(g_fp, "==== consumerId: %d, consumeMsgCnt: %" PRId64 ", consumeRowCnt: %" PRId64 "\n", pInfo->consumerId, pInfo->consumeMsgCnt, pInfo->consumeRowCnt); - + taosFsyncFile(pInfo->pConsumeRowsFile); taosCloseFile(&pInfo->pConsumeRowsFile); } @@ -882,8 +871,8 @@ void* consumeThreadFunc(void* param) { build_consumer(pInfo); build_topic_list(pInfo); if ((NULL == pInfo->tmq) || (NULL == pInfo->topicList)) { - taosFprintfFile(g_fp, "create consumer fail! tmq is null or topicList is null\n"); - taos_close(pInfo->taos); + taosFprintfFile(g_fp, "create consumer fail! tmq is null or topicList is null\n"); + taos_close(pInfo->taos); pInfo->taos = NULL; return NULL; } @@ -892,7 +881,7 @@ void* consumeThreadFunc(void* param) { if (err != 0) { pError("tmq_subscribe() fail, reason: %s\n", tmq_err2str(err)); taosFprintfFile(g_fp, "tmq_subscribe() fail! reason: %s\n", tmq_err2str(err)); - taos_close(pInfo->taos); + taos_close(pInfo->taos); pInfo->taos = NULL; return NULL; } @@ -947,7 +936,8 @@ void parseConsumeInfo() { token = strtok(g_stConfInfo.stThreads[i].topicString, delim); while (token != NULL) { // printf("%s\n", token ); - tstrncpy(g_stConfInfo.stThreads[i].topics[g_stConfInfo.stThreads[i].numOfTopic], token, sizeof(g_stConfInfo.stThreads[i].topics[g_stConfInfo.stThreads[i].numOfTopic])); + tstrncpy(g_stConfInfo.stThreads[i].topics[g_stConfInfo.stThreads[i].numOfTopic], token, + sizeof(g_stConfInfo.stThreads[i].topics[g_stConfInfo.stThreads[i].numOfTopic])); ltrim(g_stConfInfo.stThreads[i].topics[g_stConfInfo.stThreads[i].numOfTopic]); // printf("%s\n", g_stConfInfo.topics[g_stConfInfo.numOfTopic]); g_stConfInfo.stThreads[i].numOfTopic++; @@ -963,7 +953,8 @@ void parseConsumeInfo() { ltrim(pstr); char* ret = strchr(pstr, ch); memcpy(g_stConfInfo.stThreads[i].key[g_stConfInfo.stThreads[i].numOfKey], pstr, ret - pstr); - tstrncpy(g_stConfInfo.stThreads[i].value[g_stConfInfo.stThreads[i].numOfKey], ret + 1, sizeof(g_stConfInfo.stThreads[i].value[g_stConfInfo.stThreads[i].numOfKey])); + tstrncpy(g_stConfInfo.stThreads[i].value[g_stConfInfo.stThreads[i].numOfKey], ret + 1, + sizeof(g_stConfInfo.stThreads[i].value[g_stConfInfo.stThreads[i].numOfKey])); // printf("key: %s, value: %s\n", g_stConfInfo.key[g_stConfInfo.numOfKey], // g_stConfInfo.value[g_stConfInfo.numOfKey]); g_stConfInfo.stThreads[i].numOfKey++; @@ -984,12 +975,12 @@ int32_t getConsumeInfo() { } sprintf(sqlStr, "select * from %s.consumeinfo", g_stConfInfo.cdbName); - TAOS_RES *pRes = taos_query(pConn, sqlStr); + TAOS_RES* pRes = taos_query(pConn, sqlStr); if (taos_errno(pRes) != 0) { taosFprintfFile(g_fp, "error in get consumeinfo for %s\n", taos_errstr(pRes)); taosCloseFile(&g_fp); taos_free_result(pRes); - taos_close(pConn); + taos_close(pConn); return -1; } @@ -1040,19 +1031,18 @@ int32_t getConsumeInfo() { return 0; } - static int32_t omb_data_msg_process(TAOS_RES* msg, SThreadInfo* pInfo, int32_t msgIndex, int64_t* lenOfRows) { - char buf[16*1024]; + char buf[16 * 1024]; int32_t totalRows = 0; int32_t totalLen = 0; // printf("topic: %s\n", tmq_get_topic_name(msg)); - //int32_t vgroupId = tmq_get_vgroup_id(msg); - //const char* dbName = tmq_get_db_name(msg); + // int32_t vgroupId = tmq_get_vgroup_id(msg); + // const char* dbName = tmq_get_db_name(msg); - //taosFprintfFile(g_fp, "consumerId: %d, msg index:%" PRId64 "\n", pInfo->consumerId, msgIndex); - //taosFprintfFile(g_fp, "dbName: %s, topic: %s, vgroupId: %d\n", dbName != NULL ? dbName : "invalid table", - // tmq_get_topic_name(msg), vgroupId); + // taosFprintfFile(g_fp, "consumerId: %d, msg index:%" PRId64 "\n", pInfo->consumerId, msgIndex); + // taosFprintfFile(g_fp, "dbName: %s, topic: %s, vgroupId: %d\n", dbName != NULL ? dbName : "invalid table", + // tmq_get_topic_name(msg), vgroupId); while (1) { TAOS_ROW row = taos_fetch_row(msg); @@ -1061,9 +1051,9 @@ static int32_t omb_data_msg_process(TAOS_RES* msg, SThreadInfo* pInfo, int32_t m TAOS_FIELD* fields = taos_fetch_fields(msg); int32_t numOfFields = taos_field_count(msg); - //int32_t* length = taos_fetch_lengths(msg); - //int32_t precision = taos_result_precision(msg); - //const char* tbName = tmq_get_table_name(msg); + // int32_t* length = taos_fetch_lengths(msg); + // int32_t precision = taos_result_precision(msg); + // const char* tbName = tmq_get_table_name(msg); taos_print_row(buf, row, fields, numOfFields); totalLen += strlen(buf); @@ -1085,8 +1075,7 @@ void omb_loop_consume(SThreadInfo* pInfo) { char tmpString[128]; taosFprintfFile(g_fp, "%s consumer id %d start to loop pull msg\n", getCurrentTimeString(tmpString), pInfo->consumerId); - printf("%s consumer id %d start to loop pull msg\n", getCurrentTimeString(tmpString), - pInfo->consumerId); + printf("%s consumer id %d start to loop pull msg\n", getCurrentTimeString(tmpString), pInfo->consumerId); pInfo->ts = taosGetTimestampMs(); @@ -1094,55 +1083,55 @@ void omb_loop_consume(SThreadInfo* pInfo) { uint64_t lastPrintTime = taosGetTimestampMs(); uint64_t startTs = taosGetTimestampMs(); - int64_t totalLenOfMsg = 0; - int64_t lastTotalLenOfMsg = 0; + int64_t totalLenOfMsg = 0; + int64_t lastTotalLenOfMsg = 0; int32_t consumeDelay = g_stConfInfo.consumeDelay == -1 ? -1 : (g_stConfInfo.consumeDelay * 1000); while (running) { TAOS_RES* tmqMsg = tmq_consumer_poll(pInfo->tmq, consumeDelay); if (tmqMsg) { - int64_t lenOfMsg = 0; + int64_t lenOfMsg = 0; totalRows += omb_data_msg_process(tmqMsg, pInfo, totalMsgs, &lenOfMsg); - totalLenOfMsg += lenOfMsg; + totalLenOfMsg += lenOfMsg; taos_free_result(tmqMsg); totalMsgs++; int64_t currentPrintTime = taosGetTimestampMs(); if (currentPrintTime - lastPrintTime > 10 * 1000) { - int64_t currentLenOfMsg = totalLenOfMsg - lastTotalLenOfMsg; - int64_t deltaTime = currentPrintTime - lastPrintTime; - printf("consumer id %d has currently cons total rows: %" PRId64 ", msgs: %" PRId64 ", rate: %.3f msgs/s, %.1f MB/s\n", - pInfo->consumerId, totalRows, totalMsgs, - (totalMsgs - lastTotalMsgs) * 1000.0 / deltaTime, - currentLenOfMsg*1000.0/(1024*1024)/deltaTime); + int64_t currentLenOfMsg = totalLenOfMsg - lastTotalLenOfMsg; + int64_t deltaTime = currentPrintTime - lastPrintTime; + printf("consumer id %d has currently cons total rows: %" PRId64 ", msgs: %" PRId64 + ", rate: %.3f msgs/s, %.1f MB/s\n", + pInfo->consumerId, totalRows, totalMsgs, (totalMsgs - lastTotalMsgs) * 1000.0 / deltaTime, + currentLenOfMsg * 1000.0 / (1024 * 1024) / deltaTime); - taosFprintfFile( - g_fp, "consumer id %d has currently poll total msgs: %" PRId64 ", period cons rate: %.3f msgs/s, %.1f MB/s\n", - pInfo->consumerId, totalMsgs, (totalMsgs - lastTotalMsgs) * 1000.0 / deltaTime, currentLenOfMsg*1000.0/deltaTime); + taosFprintfFile(g_fp, + "consumer id %d has currently poll total msgs: %" PRId64 + ", period cons rate: %.3f msgs/s, %.1f MB/s\n", + pInfo->consumerId, totalMsgs, (totalMsgs - lastTotalMsgs) * 1000.0 / deltaTime, + currentLenOfMsg * 1000.0 / deltaTime); lastPrintTime = currentPrintTime; lastTotalMsgs = totalMsgs; - lastTotalLenOfMsg = totalLenOfMsg; + lastTotalLenOfMsg = totalLenOfMsg; } } else { char tmpString[128]; taosFprintfFile(g_fp, "%s no poll more msg when time over, break consume\n", getCurrentTimeString(tmpString)); - printf("%s no poll more msg when time over, break consume\n", getCurrentTimeString(tmpString)); + printf("%s no poll more msg when time over, break consume\n", getCurrentTimeString(tmpString)); int64_t currentPrintTime = taosGetTimestampMs(); int64_t currentLenOfMsg = totalLenOfMsg - lastTotalLenOfMsg; - int64_t deltaTime = currentPrintTime - lastPrintTime; - printf("consumer id %d has currently cons total rows: %" PRId64 ", msgs: %" PRId64 ", rate: %.3f msgs/s, %.1f MB/s\n", - pInfo->consumerId, totalRows, totalMsgs, - (totalMsgs - lastTotalMsgs) * 1000.0 / deltaTime, - currentLenOfMsg*1000.0/(1024*1024)/deltaTime); + int64_t deltaTime = currentPrintTime - lastPrintTime; + printf("consumer id %d has currently cons total rows: %" PRId64 ", msgs: %" PRId64 + ", rate: %.3f msgs/s, %.1f MB/s\n", + pInfo->consumerId, totalRows, totalMsgs, (totalMsgs - lastTotalMsgs) * 1000.0 / deltaTime, + currentLenOfMsg * 1000.0 / (1024 * 1024) / deltaTime); break; } } pInfo->consumeMsgCnt = totalMsgs; pInfo->consumeRowCnt = totalRows; - pInfo->consumeLen = totalLenOfMsg; - + pInfo->consumeLen = totalLenOfMsg; } - void* ombConsumeThreadFunc(void* param) { SThreadInfo* pInfo = (SThreadInfo*)param; @@ -1209,26 +1198,24 @@ void* ombConsumeThreadFunc(void* param) { return NULL; } +static int queryDbExec(TAOS* taos, char* command, QUERY_TYPE type) { + TAOS_RES* res = taos_query(taos, command); + int32_t code = taos_errno(res); - -static int queryDbExec(TAOS *taos, char *command, QUERY_TYPE type) { - TAOS_RES *res = taos_query(taos, command); - int32_t code = taos_errno(res); - - if (code != 0) { - pPrint("%s Failed to execute <%s>, reason: %s %s", GREEN, command, taos_errstr(res), NC); - taos_free_result(res); - return -1; - } - - if (INSERT_TYPE == type) { - int affectedRows = taos_affected_rows(res); - taos_free_result(res); - return affectedRows; - } - + if (code != 0) { + pPrint("%s Failed to execute <%s>, reason: %s %s", GREEN, command, taos_errstr(res), NC); taos_free_result(res); - return 0; + return -1; + } + + if (INSERT_TYPE == type) { + int affectedRows = taos_affected_rows(res); + taos_free_result(res); + return affectedRows; + } + + taos_free_result(res); + return 0; } void* ombProduceThreadFunc(void* param) { @@ -1236,101 +1223,100 @@ void* ombProduceThreadFunc(void* param) { pInfo->taos = createNewTaosConnect(); if (pInfo->taos == NULL) { - taosFprintfFile(g_fp, "taos_connect() fail, can not start producers!\n"); + taosFprintfFile(g_fp, "taos_connect() fail, can not start producers!\n"); return NULL; } int64_t affectedRowsTotal = 0; int64_t sendMsgs = 0; - uint32_t totalSendLoopTimes = g_stConfInfo.runDurationMinutes * 60 * 1000 / SEND_TIME_UNIT; // send some msgs per 10ms - uint32_t batchPerTblTimes = pInfo->producerRate / 100 / g_stConfInfo.batchSize; - uint32_t remainder = (pInfo->producerRate / 100) % g_stConfInfo.batchSize; + uint32_t totalSendLoopTimes = + g_stConfInfo.runDurationMinutes * 60 * 1000 / SEND_TIME_UNIT; // send some msgs per 10ms + uint32_t batchPerTblTimes = pInfo->producerRate / 100 / g_stConfInfo.batchSize; + uint32_t remainder = (pInfo->producerRate / 100) % g_stConfInfo.batchSize; if (remainder) { - batchPerTblTimes += 1; + batchPerTblTimes += 1; } char* sqlBuf = taosMemoryMalloc(MAX_SQL_LEN); if (NULL == sqlBuf) { printf("malloc fail for sqlBuf\n"); - taos_close(pInfo->taos); + taos_close(pInfo->taos); pInfo->taos = NULL; - return NULL; + return NULL; } - printf("Produce Info: totalSendLoopTimes: %d, batchPerTblTimes: %d, producerRate: %d\n", totalSendLoopTimes, batchPerTblTimes, pInfo->producerRate); + printf("Produce Info: totalSendLoopTimes: %d, batchPerTblTimes: %d, producerRate: %d\n", totalSendLoopTimes, + batchPerTblTimes, pInfo->producerRate); char ctbName[128] = {0}; sprintf(ctbName, "%s.ctb%d", g_stConfInfo.dbName, pInfo->consumerId); - int64_t lastPrintTime = taosGetTimestampUs(); - int64_t totalMsgLen = 0; - //int64_t timeStamp = taosGetTimestampUs(); + int64_t lastPrintTime = taosGetTimestampUs(); + int64_t totalMsgLen = 0; + // int64_t timeStamp = taosGetTimestampUs(); while (totalSendLoopTimes) { - int64_t startTs = taosGetTimestampUs(); + int64_t startTs = taosGetTimestampUs(); for (int i = 0; i < batchPerTblTimes; ++i) { - uint32_t msgsOfSql = g_stConfInfo.batchSize; - if ((i == batchPerTblTimes - 1) && (0 != remainder)) { - msgsOfSql = remainder; - } + uint32_t msgsOfSql = g_stConfInfo.batchSize; + if ((i == batchPerTblTimes - 1) && (0 != remainder)) { + msgsOfSql = remainder; + } int len = 0; - len += snprintf(sqlBuf+len, MAX_SQL_LEN - len, "insert into %s values ", ctbName); + len += snprintf(sqlBuf + len, MAX_SQL_LEN - len, "insert into %s values ", ctbName); for (int j = 0; j < msgsOfSql; j++) { - int64_t timeStamp = taosGetTimestampNs(); - len += snprintf(sqlBuf+len, MAX_SQL_LEN - len, "(%" PRId64 ", \"%s\")", timeStamp, g_payload); - sendMsgs++; - pInfo->totalProduceMsgs++; - } + int64_t timeStamp = taosGetTimestampNs(); + len += snprintf(sqlBuf + len, MAX_SQL_LEN - len, "(%" PRId64 ", \"%s\")", timeStamp, g_payload); + sendMsgs++; + pInfo->totalProduceMsgs++; + } - totalMsgLen += len; - pInfo->totalMsgsLen += len; - - int64_t affectedRows = queryDbExec(pInfo->taos, sqlBuf, INSERT_TYPE); - if (affectedRows < 0) { - taos_close(pInfo->taos); - pInfo->taos = NULL; - taosMemoryFree(sqlBuf); - return NULL; - } + totalMsgLen += len; + pInfo->totalMsgsLen += len; - affectedRowsTotal += affectedRows; + int64_t affectedRows = queryDbExec(pInfo->taos, sqlBuf, INSERT_TYPE); + if (affectedRows < 0) { + taos_close(pInfo->taos); + pInfo->taos = NULL; + taosMemoryFree(sqlBuf); + return NULL; + } - //printf("Produce Info: affectedRows: %" PRId64 "\n", affectedRows); + affectedRowsTotal += affectedRows; + + // printf("Produce Info: affectedRows: %" PRId64 "\n", affectedRows); } totalSendLoopTimes -= 1; - // calc spent time - int64_t currentTs = taosGetTimestampUs(); - int64_t delta = currentTs - startTs; - if (delta < SEND_TIME_UNIT * 1000) { - int64_t sleepLen = (int32_t)(SEND_TIME_UNIT * 1000 - delta); - //printf("sleep %" PRId64 " us, use time: %" PRId64 " us\n", sleepLen, delta); - taosUsleep((int32_t)sleepLen); - } + // calc spent time + int64_t currentTs = taosGetTimestampUs(); + int64_t delta = currentTs - startTs; + if (delta < SEND_TIME_UNIT * 1000) { + int64_t sleepLen = (int32_t)(SEND_TIME_UNIT * 1000 - delta); + // printf("sleep %" PRId64 " us, use time: %" PRId64 " us\n", sleepLen, delta); + taosUsleep((int32_t)sleepLen); + } currentTs = taosGetTimestampUs(); delta = currentTs - lastPrintTime; - if (delta > 10 * 1000 * 1000) { - printf("producer[%d] info: %" PRId64 " msgs, %" PRId64 " Byte, %" PRId64 " us, totalSendLoopTimes: %d\n", - pInfo->consumerId, sendMsgs, totalMsgLen, delta, totalSendLoopTimes); - printf("producer[%d] rate: %1.f msgs/s, %1.f KB/s\n", - pInfo->consumerId, - sendMsgs * 1000.0 * 1000 / delta, - (totalMsgLen / 1024.0) / (delta / (1000*1000))); - lastPrintTime = currentTs; - sendMsgs = 0; - totalMsgLen = 0; - } + if (delta > 10 * 1000 * 1000) { + printf("producer[%d] info: %" PRId64 " msgs, %" PRId64 " Byte, %" PRId64 " us, totalSendLoopTimes: %d\n", + pInfo->consumerId, sendMsgs, totalMsgLen, delta, totalSendLoopTimes); + printf("producer[%d] rate: %1.f msgs/s, %1.f KB/s\n", pInfo->consumerId, sendMsgs * 1000.0 * 1000 / delta, + (totalMsgLen / 1024.0) / (delta / (1000 * 1000))); + lastPrintTime = currentTs; + sendMsgs = 0; + totalMsgLen = 0; + } } - printf("affectedRowsTotal: %"PRId64"\n", affectedRowsTotal); + printf("affectedRowsTotal: %" PRId64 "\n", affectedRowsTotal); taos_close(pInfo->taos); pInfo->taos = NULL; taosMemoryFree(sqlBuf); return NULL; } - void printProduceInfo(int64_t start) { int64_t totalMsgs = 0; int64_t totalLenOfMsgs = 0; @@ -1347,87 +1333,86 @@ void printProduceInfo(int64_t start) { double tInMs = (double)t / 1000000.0; printf("Spent %.3f seconds to prod %" PRIu64 " msgs, %" PRIu64 " Byte\n\n", tInMs, totalMsgs, totalLenOfMsgs); - printf("Spent %.3f seconds to prod %" PRIu64 " msgs with %d producer(s), throughput: %.3f msgs/s, %.1f MB/s\n\n", - tInMs, totalMsgs, g_stConfInfo.producers, - (double)totalMsgs / tInMs, - (double)totalLenOfMsgs/(1024.0*1024)/tInMs); + tInMs, totalMsgs, g_stConfInfo.producers, (double)totalMsgs / tInMs, + (double)totalLenOfMsgs / (1024.0 * 1024) / tInMs); return; } - void startOmbConsume() { - TdThreadAttr thattr; - taosThreadAttrInit(&thattr); - taosThreadAttrSetDetachState(&thattr, PTHREAD_CREATE_JOINABLE); + TdThreadAttr thattr; + taosThreadAttrInit(&thattr); + taosThreadAttrSetDetachState(&thattr, PTHREAD_CREATE_JOINABLE); if (0 != g_stConfInfo.producers) { TAOS* taos = createNewTaosConnect(); if (taos == NULL) { taosFprintfFile(g_fp, "taos_connect() fail, can not create db, stbl, ctbl, topic!\n"); - return ; + return; } - char stbName[16] = "stb"; - char ctbPrefix[16] = "ctb"; - + char stbName[16] = "stb"; + char ctbPrefix[16] = "ctb"; + char sql[256] = {0}; sprintf(sql, "drop database if exists %s", g_stConfInfo.dbName); - printf("SQL: %s\n", sql); - queryDbExec(taos, sql, NO_INSERT_TYPE); - - sprintf(sql, "create database if not exists %s precision 'ns' vgroups %d", g_stConfInfo.dbName, g_stConfInfo.producers); - printf("SQL: %s\n", sql); + printf("SQL: %s\n", sql); queryDbExec(taos, sql, NO_INSERT_TYPE); - sprintf(sql, "create stable %s.%s (ts timestamp, payload binary(%d)) tags (t bigint) ", g_stConfInfo.dbName, stbName, g_stConfInfo.payloadLen); - printf("SQL: %s\n", sql); + sprintf(sql, "create database if not exists %s precision 'ns' vgroups %d", g_stConfInfo.dbName, + g_stConfInfo.producers); + printf("SQL: %s\n", sql); queryDbExec(taos, sql, NO_INSERT_TYPE); - for (int i = 0; i < g_stConfInfo.producers; i++) { - sprintf(sql, "create table %s.%s%d using %s.stb tags(%d) ", g_stConfInfo.dbName, ctbPrefix, i, g_stConfInfo.dbName, i); - printf("SQL: %s\n", sql); + sprintf(sql, "create stable %s.%s (ts timestamp, payload binary(%d)) tags (t bigint) ", g_stConfInfo.dbName, + stbName, g_stConfInfo.payloadLen); + printf("SQL: %s\n", sql); + queryDbExec(taos, sql, NO_INSERT_TYPE); + + for (int i = 0; i < g_stConfInfo.producers; i++) { + sprintf(sql, "create table %s.%s%d using %s.stb tags(%d) ", g_stConfInfo.dbName, ctbPrefix, i, + g_stConfInfo.dbName, i); + printf("SQL: %s\n", sql); queryDbExec(taos, sql, NO_INSERT_TYPE); - } + } - // create topic + // create topic sprintf(sql, "create topic %s as stable %s.%s", g_stConfInfo.topic, g_stConfInfo.dbName, stbName); - printf("SQL: %s\n", sql); + printf("SQL: %s\n", sql); queryDbExec(taos, sql, NO_INSERT_TYPE); - - int32_t producerRate = ceil(g_stConfInfo.producerRate / g_stConfInfo.producers); - - printf("==== create %d produce thread ====\n", g_stConfInfo.producers); - for (int32_t i = 0; i < g_stConfInfo.producers; ++i) { - g_stConfInfo.stProdThreads[i].consumerId = i; - g_stConfInfo.stProdThreads[i].producerRate = producerRate; - taosThreadCreate(&(g_stConfInfo.stProdThreads[i].thread), &thattr, ombProduceThreadFunc, - (void*)(&(g_stConfInfo.stProdThreads[i]))); - } - - if (0 == g_stConfInfo.numOfThread) { - int64_t start = taosGetTimestampUs(); + int32_t producerRate = ceil(g_stConfInfo.producerRate / g_stConfInfo.producers); + + printf("==== create %d produce thread ====\n", g_stConfInfo.producers); + for (int32_t i = 0; i < g_stConfInfo.producers; ++i) { + g_stConfInfo.stProdThreads[i].consumerId = i; + g_stConfInfo.stProdThreads[i].producerRate = producerRate; + taosThreadCreate(&(g_stConfInfo.stProdThreads[i].thread), &thattr, ombProduceThreadFunc, + (void*)(&(g_stConfInfo.stProdThreads[i]))); + } + + if (0 == g_stConfInfo.numOfThread) { + int64_t start = taosGetTimestampUs(); for (int32_t i = 0; i < g_stConfInfo.producers; i++) { taosThreadJoin(g_stConfInfo.stProdThreads[i].thread, NULL); taosThreadClear(&g_stConfInfo.stProdThreads[i].thread); } - printProduceInfo(start); - - taosFprintfFile(g_fp, "==== close tmqlog ====\n"); - taosCloseFile(&g_fp); - taos_close(taos); - return; - } + printProduceInfo(start); - taos_close(taos); + taosFprintfFile(g_fp, "==== close tmqlog ====\n"); + taosCloseFile(&g_fp); + taos_close(taos); + return; + } + + taos_close(taos); } // pthread_create one thread to consume taosFprintfFile(g_fp, "==== create %d consume thread ====\n", g_stConfInfo.numOfThread); for (int32_t i = 0; i < g_stConfInfo.numOfThread; ++i) { - g_stConfInfo.stThreads[i].consumerId = i; + g_stConfInfo.stThreads[i].consumerId = i; taosThreadCreate(&(g_stConfInfo.stThreads[i].thread), &thattr, ombConsumeThreadFunc, (void*)(&(g_stConfInfo.stThreads[i]))); } @@ -1446,24 +1431,23 @@ void startOmbConsume() { int64_t totalLenOfMsgs = 0; for (int32_t i = 0; i < g_stConfInfo.numOfThread; i++) { totalMsgs += g_stConfInfo.stThreads[i].consumeMsgCnt; - totalLenOfMsgs += g_stConfInfo.stThreads[i].consumeLen; - totalRows += g_stConfInfo.stThreads[i].consumeRowCnt; + totalLenOfMsgs += g_stConfInfo.stThreads[i].consumeLen; + totalRows += g_stConfInfo.stThreads[i].consumeRowCnt; } int64_t t = end - start; if (0 == t) t = 1; double tInMs = (double)t / 1000000.0; - taosFprintfFile(g_fp, - "Spent %.3f seconds to poll msgs: %" PRIu64 " with %d thread(s), throughput: %.3f msgs/s, %.1f MB/s\n\n", - tInMs, totalMsgs, g_stConfInfo.numOfThread, - (double)(totalMsgs / tInMs), - (double)totalLenOfMsgs/(1024*1024)/tInMs); + taosFprintfFile( + g_fp, "Spent %.3f seconds to poll msgs: %" PRIu64 " with %d thread(s), throughput: %.3f msgs/s, %.1f MB/s\n\n", + tInMs, totalMsgs, g_stConfInfo.numOfThread, (double)(totalMsgs / tInMs), + (double)totalLenOfMsgs / (1024 * 1024) / tInMs); - printf("Spent %.3f seconds to cons rows: %" PRIu64 " msgs: %" PRIu64 " with %d thread(s), throughput: %.3f msgs/s, %.1f MB/s\n\n", - tInMs, totalRows, totalMsgs, g_stConfInfo.numOfThread, - (double)(totalMsgs / tInMs), - (double)totalLenOfMsgs/(1024*1024)/tInMs); + printf("Spent %.3f seconds to cons rows: %" PRIu64 " msgs: %" PRIu64 + " with %d thread(s), throughput: %.3f msgs/s, %.1f MB/s\n\n", + tInMs, totalRows, totalMsgs, g_stConfInfo.numOfThread, (double)(totalMsgs / tInMs), + (double)totalLenOfMsgs / (1024 * 1024) / tInMs); taosFprintfFile(g_fp, "==== close tmqlog ====\n"); taosCloseFile(&g_fp); @@ -1471,20 +1455,19 @@ void startOmbConsume() { return; } - int main(int32_t argc, char* argv[]) { parseArgument(argc, argv); if (0 != strlen(g_stConfInfo.topic)) { startOmbConsume(); - return 0; + return 0; } - + int32_t retCode = getConsumeInfo(); if (0 != retCode) { return -1; } - + saveConfigToLogFile(); tmqSetSignalHandle(); From e29e477b0c94ca4efe05be3d5d8a4d2b7f5b7114 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sun, 27 Nov 2022 15:46:31 +0800 Subject: [PATCH 68/83] fix(query): update api to avoid deadlock. --- source/libs/executor/src/executor.c | 2 +- source/libs/function/src/detail/tminmax.c | 2 ++ 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/source/libs/executor/src/executor.c b/source/libs/executor/src/executor.c index 01baba5a52..9b3bd1d808 100644 --- a/source/libs/executor/src/executor.c +++ b/source/libs/executor/src/executor.c @@ -287,7 +287,7 @@ static SArray* filterUnqualifiedTables(const SStreamScanInfo* pScanInfo, const S for (int32_t i = 0; i < taosArrayGetSize(tableIdList); ++i) { uint64_t* id = (uint64_t*)taosArrayGet(tableIdList, i); - int32_t code = metaGetTableEntryByUidCache(&mr, *id); + int32_t code = metaGetTableEntryByUid(&mr, *id); if (code != TSDB_CODE_SUCCESS) { qError("failed to get table meta, uid:%" PRIu64 " code:%s, %s", *id, tstrerror(terrno), idstr); continue; diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c index e47edb8a1e..b2cb36cba0 100644 --- a/source/libs/function/src/detail/tminmax.c +++ b/source/libs/function/src/detail/tminmax.c @@ -877,6 +877,8 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { break; } } + + pBuf->assign = true; } _over: From ecff9423bf2d963700efccb7e43d3fa889811a15 Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 15:40:43 +0800 Subject: [PATCH 69/83] test: add asan case --- tests/parallel_test/cases.task | 6 +++--- tests/system-test/7-tmq/tmqCommon.py | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/tests/parallel_test/cases.task b/tests/parallel_test/cases.task index e7e273bc9f..cecaaef1d0 100644 --- a/tests/parallel_test/cases.task +++ b/tests/parallel_test/cases.task @@ -687,9 +687,9 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqError.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/schema.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/stbFilter.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqCheckData.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqCheckData1.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsumerGroup.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqCheckData.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqCheckData1.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsumerGroup.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqShow.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqAlterSchema.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb.py diff --git a/tests/system-test/7-tmq/tmqCommon.py b/tests/system-test/7-tmq/tmqCommon.py index c153e94caa..141d013270 100644 --- a/tests/system-test/7-tmq/tmqCommon.py +++ b/tests/system-test/7-tmq/tmqCommon.py @@ -128,12 +128,12 @@ class TMQCom: os.system(shellCmd) def stopTmqSimProcess(self, processorName): - psCmd = "ps -ef|grep -w %s|grep -v grep | awk '{print $2}'"%(processorName) + psCmd = "unset LD_PRELOAD; ps -ef|grep -w %s|grep -v grep | awk '{print $2}'"%(processorName) processID = subprocess.check_output(psCmd, shell=True).decode("utf-8") onlyKillOnceWindows = 0 while(processID): if not platform.system().lower() == 'windows' or (onlyKillOnceWindows == 0 and platform.system().lower() == 'windows'): - killCmd = "kill -INT %s > /dev/null 2>&1" % processID + killCmd = "unset LD_PRELOAD; kill -INT %s > /dev/null 2>&1" % processID os.system(killCmd) onlyKillOnceWindows = 1 time.sleep(0.2) From a8a75f33d7ad5d382e033f69543458627b5d0d0e Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 16:15:58 +0800 Subject: [PATCH 70/83] test: adjust tmqshow case --- tests/parallel_test/cases.task | 2 +- tests/system-test/7-tmq/tmqShow.py | 23 +++++++++++++++-------- 2 files changed, 16 insertions(+), 9 deletions(-) diff --git a/tests/parallel_test/cases.task b/tests/parallel_test/cases.task index cecaaef1d0..991e2fecd6 100644 --- a/tests/parallel_test/cases.task +++ b/tests/parallel_test/cases.task @@ -690,7 +690,7 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqCheckData.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqCheckData1.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsumerGroup.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqShow.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqShow.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqAlterSchema.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1.py diff --git a/tests/system-test/7-tmq/tmqShow.py b/tests/system-test/7-tmq/tmqShow.py index 8ec5e62ad9..31ddc1b0f8 100644 --- a/tests/system-test/7-tmq/tmqShow.py +++ b/tests/system-test/7-tmq/tmqShow.py @@ -126,14 +126,21 @@ class TDTestCase: pThread = tmqCom.asyncInsertData(paraDict) tmqCom.getStartConsumeNotifyFromTmqsim() - #time.sleep(5) - tdLog.info("check show consumers") - tdSql.query("show consumers") - # tdLog.info(tdSql.queryResult) - rows = tdSql.getRows() - tdLog.info("show consumers rows: %d"%rows) - if rows != len(topicNameList): - tdLog.exit("show consumers rows error") + + for i in range(0, 10, 1): + tdLog.info("check show consumers") + tdSql.query("show consumers") + # tdLog.info(tdSql.queryResult) + rows = tdSql.getRows() + tdLog.info("show consumers rows: %d" % rows) + + if rows == len(topicNameList): + tdLog.info("show consumers rows not match %d:%d" % + (rows, len(topicNameList))) + time.sleep(1) + break + if (rows == 9): + tdLog.exit("show consumers rows error") for i in range(0, 10, 1): tdLog.info("check show subscriptions") From 2dc23cdfe559d612872dc16d2848b29b5e313cbc Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 17:49:23 +0800 Subject: [PATCH 71/83] test: adjust tmqCommon cases --- tests/parallel_test/cases.task | 16 ++++++++-------- tests/pytest/util/common.py | 2 +- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/tests/parallel_test/cases.task b/tests/parallel_test/cases.task index 991e2fecd6..6eb3fb29f2 100644 --- a/tests/parallel_test/cases.task +++ b/tests/parallel_test/cases.task @@ -692,14 +692,14 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsumerGroup.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqShow.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqAlterSchema.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb-1ctb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb-1ctb-funcNFilter.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb-funcNFilter.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-1ctb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-1ctb-funcNFilter.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb-funcNFilter.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb-funcNFilter.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb-funcNFilter.py diff --git a/tests/pytest/util/common.py b/tests/pytest/util/common.py index 9ffebcbdad..5b73989d6f 100644 --- a/tests/pytest/util/common.py +++ b/tests/pytest/util/common.py @@ -737,7 +737,7 @@ class TDCom: if (platform.system().lower() == 'windows'): os.system("TASKKILL /F /IM %s.exe"%processorName) else: - os.system('pkill %s'%processorName) + os.system("unset LD_PRELOAD; pkill %s " % processorName) def is_json(msg): From 1e8dcacea6347367315d5d2bacd96828bb793815 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Sun, 27 Nov 2022 18:25:20 +0800 Subject: [PATCH 72/83] fix(query): set null for inf and nan value. --- source/libs/function/src/detail/tavgfunction.c | 7 ++++++- tests/system-test/2-query/avg.py | 6 +++--- 2 files changed, 9 insertions(+), 4 deletions(-) diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c index 267cb36769..7d018a8dc7 100644 --- a/source/libs/function/src/detail/tavgfunction.c +++ b/source/libs/function/src/detail/tavgfunction.c @@ -723,7 +723,12 @@ int32_t avgFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { } } - pEntryInfo->numOfRes = (pRes->count > 0)? 1:0; + if (pRes->count == 0 || isinf(pRes->result) || isnan(pRes->result)) { + pEntryInfo->numOfRes = 0; + } else { + pEntryInfo->numOfRes = 1; + } + return functionFinalize(pCtx, pBlock); } diff --git a/tests/system-test/2-query/avg.py b/tests/system-test/2-query/avg.py index 139e7d4bf4..1d4d9a2494 100644 --- a/tests/system-test/2-query/avg.py +++ b/tests/system-test/2-query/avg.py @@ -15,7 +15,7 @@ class TDTestCase: def init(self, conn, logSql, replicaVar=1): self.replicaVar = int(replicaVar) tdLog.debug(f"start to excute {__file__}") - tdSql.init(conn.cursor(), False) + tdSql.init(conn.cursor(), True) self.setsql = TDSetSql() self.column_dict = { 'ts':'timestamp', @@ -413,7 +413,7 @@ class TDTestCase: tdSql.checkData(0,2,14042.142857143) tdSql.checkData(0,3,53.571428571) tdSql.checkData(0,4,5.828571332045761e+37) - tdSql.checkData(0,5,math.inf) + tdSql.checkData(0,5,None) # check + - * / in functions @@ -423,7 +423,7 @@ class TDTestCase: tdSql.checkData(0,2,14042.142857143) tdSql.checkData(0,3,26.785714286) tdSql.checkData(0,4,2.9142856660228804e+37) - tdSql.checkData(0,5,math.inf) + tdSql.checkData(0,5,None) From 56ccb65c3e174c31cd104aabe99e93b563045e2a Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 20:57:58 +0800 Subject: [PATCH 73/83] test: add asan case --- tests/parallel_test/cases.task | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tests/parallel_test/cases.task b/tests/parallel_test/cases.task index 6eb3fb29f2..673def27e8 100644 --- a/tests/parallel_test/cases.task +++ b/tests/parallel_test/cases.task @@ -700,12 +700,12 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-1ctb-funcNFilter.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb-funcNFilter.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb-funcNFilter.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb-funcNFilter.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb-funcNFilter.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqAutoCreateTbl.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqDnodeRestart.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqDnodeRestart.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqDnodeRestart1.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqUpdate-1ctb.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqUpdateWithConsume.py From c4045de4bb212aa5903435292786b293566434a4 Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 21:27:15 +0800 Subject: [PATCH 74/83] fix: memory leak while subscribe --- source/dnode/mnode/impl/src/mndScheduler.c | 2 ++ source/dnode/mnode/impl/src/mndTopic.c | 2 ++ 2 files changed, 4 insertions(+) diff --git a/source/dnode/mnode/impl/src/mndScheduler.c b/source/dnode/mnode/impl/src/mndScheduler.c index 3c1d3f09bf..f71bd1c626 100644 --- a/source/dnode/mnode/impl/src/mndScheduler.c +++ b/source/dnode/mnode/impl/src/mndScheduler.c @@ -585,6 +585,8 @@ int32_t mndSchedInitSubEp(SMnode* pMnode, const SMqTopicObj* pTopic, SMqSubscrib } else { pVgEp->qmsg = strdup(""); } + + sdbRelease(pSdb, pVgroup); } ASSERT(pSub->unassignedVgs->size > 0); diff --git a/source/dnode/mnode/impl/src/mndTopic.c b/source/dnode/mnode/impl/src/mndTopic.c index 522036afa2..f1db6a12fc 100644 --- a/source/dnode/mnode/impl/src/mndTopic.c +++ b/source/dnode/mnode/impl/src/mndTopic.c @@ -512,6 +512,8 @@ static int32_t mndCreateTopic(SMnode *pMnode, SRpcMsg *pReq, SCMCreateTopicReq * mndTransDrop(pTrans); return -1; } + + sdbRelease(pSdb, pVgroup); } } From 7112a4ca6cb8edca44e1c060aa68a023ebca5430 Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 21:40:16 +0800 Subject: [PATCH 75/83] test: add asan case --- tests/parallel_test/cases.task | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/tests/parallel_test/cases.task b/tests/parallel_test/cases.task index 673def27e8..233f3fbc6e 100644 --- a/tests/parallel_test/cases.task +++ b/tests/parallel_test/cases.task @@ -691,7 +691,7 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqCheckData1.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsumerGroup.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqShow.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqAlterSchema.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqAlterSchema.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg.py @@ -704,20 +704,20 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb-funcNFilter.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb-funcNFilter.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqAutoCreateTbl.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqAutoCreateTbl.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqDnodeRestart.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqDnodeRestart1.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqUpdate-1ctb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqUpdateWithConsume.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUpdate-1ctb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUpdateWithConsume.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUpdate-multiCtb-snapshot0.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUpdate-multiCtb-snapshot1.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqDelete-1ctb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqDelete-multiCtb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqDelete-1ctb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqDelete-multiCtb.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqDropStb.py ,,,system-test,python3 ./test.py -f 7-tmq/tmqDropStbCtb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqDropNtb-snapshot0.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqDropNtb-snapshot1.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqUdf.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqDropNtb-snapshot0.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqDropNtb-snapshot1.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUdf.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUdf-multCtb-snapshot0.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUdf-multCtb-snapshot1.py ,,,system-test,python3 ./test.py -f 7-tmq/stbTagFilter-1ctb.py From e4805a0cee405a5ba64c6f3c56b5d6a547b74ded Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 21:50:54 +0800 Subject: [PATCH 76/83] test: add asan case --- tests/parallel_test/cases.task | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tests/parallel_test/cases.task b/tests/parallel_test/cases.task index 233f3fbc6e..f99005af2d 100644 --- a/tests/parallel_test/cases.task +++ b/tests/parallel_test/cases.task @@ -720,11 +720,11 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUdf.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUdf-multCtb-snapshot0.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqUdf-multCtb-snapshot1.py -,,,system-test,python3 ./test.py -f 7-tmq/stbTagFilter-1ctb.py -,,,system-test,python3 ./test.py -f 7-tmq/dataFromTsdbNWal.py -,,,system-test,python3 ./test.py -f 7-tmq/dataFromTsdbNWal-multiCtb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/stbTagFilter-1ctb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/dataFromTsdbNWal.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/dataFromTsdbNWal-multiCtb.py ,,,system-test,python3 ./test.py -f 7-tmq/tmq_taosx.py -,,,system-test,python3 ./test.py -f 7-tmq/stbTagFilter-multiCtb.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/stbTagFilter-multiCtb.py ,,y,system-test,./pytest.sh python3 ./test.py -f 99-TDcase/TD-19201.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqSubscribeStb-r3.py -N 5 ,,,system-test,python3 ./test.py -f 7-tmq/tmq3mnodeSwitch.py -N 6 -M 3 From 74f83a362bb85d241054ffc222dbb8286e1e98a6 Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 22:23:40 +0800 Subject: [PATCH 77/83] fix: memory while use taosx --- source/dnode/mnode/impl/src/mndStream.c | 1 + source/dnode/mnode/impl/src/mndTopic.c | 1 + 2 files changed, 2 insertions(+) diff --git a/source/dnode/mnode/impl/src/mndStream.c b/source/dnode/mnode/impl/src/mndStream.c index 62247f2c2a..d8cf7a837e 100644 --- a/source/dnode/mnode/impl/src/mndStream.c +++ b/source/dnode/mnode/impl/src/mndStream.c @@ -525,6 +525,7 @@ static int32_t mndCreateStbForStream(SMnode *pMnode, STrans *pTrans, const SStre tFreeSMCreateStbReq(&createReq); mndFreeStb(&stbObj); + mndReleaseStb(pMnode, pStb); mndReleaseDb(pMnode, pDb); return 0; diff --git a/source/dnode/mnode/impl/src/mndTopic.c b/source/dnode/mnode/impl/src/mndTopic.c index f1db6a12fc..6412761f0b 100644 --- a/source/dnode/mnode/impl/src/mndTopic.c +++ b/source/dnode/mnode/impl/src/mndTopic.c @@ -434,6 +434,7 @@ static int32_t mndCreateTopic(SMnode *pMnode, SRpcMsg *pReq, SCMCreateTopicReq * return -1; } topicObj.stbUid = pStb->uid; + mndReleaseStb(pMnode, pStb); } /*} else if (pCreate->subType == TOPIC_SUB_TYPE__DB) {*/ /*topicObj.ast = NULL;*/ From fd65b8935b34ee3cb080c9838d8871e45d82368e Mon Sep 17 00:00:00 2001 From: Shengliang Guan Date: Sun, 27 Nov 2022 22:24:28 +0800 Subject: [PATCH 78/83] test: add asan case --- tests/parallel_test/cases.task | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tests/parallel_test/cases.task b/tests/parallel_test/cases.task index f99005af2d..1b554f928c 100644 --- a/tests/parallel_test/cases.task +++ b/tests/parallel_test/cases.task @@ -702,7 +702,7 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb-funcNFilter.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb-mutilVg-mutilCtb.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-1ctb-funcNFilter.py -,,,system-test,python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb-funcNFilter.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb-funcNFilter.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqConsFromTsdb1-mutilVg-mutilCtb.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqAutoCreateTbl.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqDnodeRestart.py @@ -723,12 +723,12 @@ ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/stbTagFilter-1ctb.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/dataFromTsdbNWal.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/dataFromTsdbNWal-multiCtb.py -,,,system-test,python3 ./test.py -f 7-tmq/tmq_taosx.py +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmq_taosx.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/stbTagFilter-multiCtb.py ,,y,system-test,./pytest.sh python3 ./test.py -f 99-TDcase/TD-19201.py ,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmqSubscribeStb-r3.py -N 5 -,,,system-test,python3 ./test.py -f 7-tmq/tmq3mnodeSwitch.py -N 6 -M 3 -,,,system-test,python3 ./test.py -f 7-tmq/tmq3mnodeSwitch.py -N 6 -M 3 -n 3 +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmq3mnodeSwitch.py -N 6 -M 3 +,,y,system-test,./pytest.sh python3 ./test.py -f 7-tmq/tmq3mnodeSwitch.py -N 6 -M 3 -n 3 ,,y,system-test,./pytest.sh python3 ./test.py -f 2-query/between.py -Q 2 ,,y,system-test,./pytest.sh python3 ./test.py -f 2-query/distinct.py -Q 2 ,,y,system-test,./pytest.sh python3 ./test.py -f 2-query/varchar.py -Q 2 From e8764790aaf10691541749da7ef02e929dc36e9d Mon Sep 17 00:00:00 2001 From: Shuduo Sang Date: Sun, 27 Nov 2022 22:55:34 +0800 Subject: [PATCH 79/83] feat: taosbenchmark supports retry (#18496) --- cmake/taostools_CMakeLists.txt.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/cmake/taostools_CMakeLists.txt.in b/cmake/taostools_CMakeLists.txt.in index 30e7b85495..577166353b 100644 --- a/cmake/taostools_CMakeLists.txt.in +++ b/cmake/taostools_CMakeLists.txt.in @@ -2,7 +2,7 @@ # taos-tools ExternalProject_Add(taos-tools GIT_REPOSITORY https://github.com/taosdata/taos-tools.git - GIT_TAG 4cbb9ac + GIT_TAG 7e9ce09 SOURCE_DIR "${TD_SOURCE_DIR}/tools/taos-tools" BINARY_DIR "" #BUILD_IN_SOURCE TRUE From 1e8323fac80f00aca9ede59f6837584678aa8820 Mon Sep 17 00:00:00 2001 From: Shuduo Sang Date: Mon, 28 Nov 2022 00:23:52 +0800 Subject: [PATCH 80/83] docs: update examples/jdbc demo readme (#18499) * docs: update csharp connector status * docs: fix csharp ws bulk pulling * docs: clarify database param is optional to websocket dsn * docs: fix python version and a few typos * docs: fix jdbc version in connector matrix * docs: update jdbc demo readme --- examples/JDBC/taosdemo/pom.xml | 2 +- examples/JDBC/taosdemo/readme.md | 10 ++++------ 2 files changed, 5 insertions(+), 7 deletions(-) diff --git a/examples/JDBC/taosdemo/pom.xml b/examples/JDBC/taosdemo/pom.xml index 724ecc7407..68224bbad5 100644 --- a/examples/JDBC/taosdemo/pom.xml +++ b/examples/JDBC/taosdemo/pom.xml @@ -88,7 +88,7 @@ org.apache.logging.log4j log4j-core - 2.17.1 + 2.17.2 diff --git a/examples/JDBC/taosdemo/readme.md b/examples/JDBC/taosdemo/readme.md index e5f4eb132b..edac970399 100644 --- a/examples/JDBC/taosdemo/readme.md +++ b/examples/JDBC/taosdemo/readme.md @@ -2,12 +2,10 @@ cd tests/examples/JDBC/taosdemo mvn clean package -Dmaven.test.skip=true # 先建表，再插入的 -java -jar target/taosdemo-2.0.1-jar-with-dependencies.jar -host [hostname] -database [database] -doCreateTable true -superTableSQL "create table weather(ts timestamp, f1 int) tags(t1 nchar(4))" -numOfTables 1000 -numOfRowsPerTable 100000000 -numOfThreadsForInsert 10 -numOfTablesPerSQL 10 -numOfValuesPerSQL 100 +java -jar target/taosdemo-2.0.1-jar-with-dependencies.jar -host -database -doCreateTable true -superTableSQL "create table weather(ts timestamp, f1 int) tags(t1 nchar(4))" -numOfTables 1000 -numOfRowsPerTable 100000000 -numOfThreadsForInsert 10 -numOfTablesPerSQL 10 -numOfValuesPerSQL 100 # 不建表，直接插入的 -java -jar target/taosdemo-2.0.1-jar-with-dependencies.jar -host [hostname] -database [database] -doCreateTable false -superTableSQL "create table weather(ts timestamp, f1 int) tags(t1 nchar(4))" -numOfTables 1000 -numOfRowsPerTable 100000000 -numOfThreadsForInsert 10 -numOfTablesPerSQL 10 -numOfValuesPerSQL 100 +java -jar target/taosdemo-2.0.1-jar-with-dependencies.jar -host -database -doCreateTable false -superTableSQL "create table weather(ts timestamp, f1 int) tags(t1 nchar(4))" -numOfTables 1000 -numOfRowsPerTable 100000000 -numOfThreadsForInsert 10 -numOfTablesPerSQL 10 -numOfValuesPerSQL 100 ``` -需求： -1. 可以读lowa的配置文件 -2. 支持JDBC-JNI和JDBC-restful -3. 读取配置文件，持续执行查询 \ No newline at end of file +如果发生错误 Exception in thread "main" java.lang.UnsatisfiedLinkError: no taos in java.library.path +请检查是否安装 TDengine 客户端安装包或编译 TDengine 安装。如果确定已经安装过还出现这个错误，可以在命令行 java 后加 -Djava.library.path=/usr/local/lib 来指定寻找共享库的路径。 From b0ee829db25f0ee6921c0359ddb4b25e74fbb16a Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Mon, 28 Nov 2022 00:27:49 +0800 Subject: [PATCH 81/83] refactor: do some internal refactor. --- source/libs/executor/inc/executil.h | 2 +- source/libs/executor/inc/executorimpl.h | 159 +--- source/libs/executor/src/cachescanoperator.c | 2 +- source/libs/executor/src/exchangeoperator.c | 2 +- source/libs/executor/src/executil.c | 2 +- source/libs/executor/src/executorimpl.c | 19 +- source/libs/executor/src/groupoperator.c | 10 +- source/libs/executor/src/joinoperator.c | 2 +- source/libs/executor/src/projectoperator.c | 4 +- source/libs/executor/src/scanoperator.c | 15 +- source/libs/executor/src/sortoperator.c | 8 +- source/libs/executor/src/sysscanoperator.c | 4 +- source/libs/executor/src/tfill.c | 6 +- source/libs/executor/src/timewindowoperator.c | 618 +----------- source/libs/function/src/builtinsimpl.c | 893 ------------------ 15 files changed, 141 insertions(+), 1605 deletions(-) diff --git a/source/libs/executor/inc/executil.h b/source/libs/executor/inc/executil.h index fd8a357a8b..d5366f1b7a 100644 --- a/source/libs/executor/inc/executil.h +++ b/source/libs/executor/inc/executil.h @@ -140,7 +140,7 @@ bool hasRemainResults(SGroupResInfo* pGroupResInfo); int32_t getNumOfTotalRes(SGroupResInfo* pGroupResInfo); -SSDataBlock* createResDataBlock(SDataBlockDescNode* pNode); +SSDataBlock* createDataBlockFromDescNode(SDataBlockDescNode* pNode); EDealRes doTranslateTagExpr(SNode** pNode, void* pContext); int32_t getGroupIdFromTagsVal(void* pMeta, uint64_t uid, SNodeList* pGroupNode, char* keyBuf, uint64_t* pGroupId); diff --git a/source/libs/executor/inc/executorimpl.h b/source/libs/executor/inc/executorimpl.h index b9d622a4c3..bd4472327c 100644 --- a/source/libs/executor/inc/executorimpl.h +++ b/source/libs/executor/inc/executorimpl.h @@ -537,15 +537,6 @@ typedef struct SStreamIntervalOperatorInfo { SWinKey delKey; } SStreamIntervalOperatorInfo; -typedef struct SAggOperatorInfo { - SOptrBasicInfo binfo; - SAggSupporter aggSup; - STableQueryInfo* current; - uint64_t groupId; - SGroupResInfo groupResInfo; - SExprSupp scalarExprSup; -} SAggOperatorInfo; - typedef struct SFillOperatorInfo { struct SFillInfo* pFillInfo; SSDataBlock* pRes; @@ -577,18 +568,6 @@ typedef struct SWindowRowsSup { uint64_t groupId; } SWindowRowsSup; -typedef struct SSessionAggOperatorInfo { - SOptrBasicInfo binfo; - SAggSupporter aggSup; - - SGroupResInfo groupResInfo; - SWindowRowsSup winSup; - bool reptScan; // next round scan - int64_t gap; // session window gap - int32_t tsSlotId; // primary timestamp slot id - STimeWindowAggSupp twAggSup; -} SSessionAggOperatorInfo; - typedef struct SResultWindowInfo { void* pOutputBuf; SSessionKey sessionWin; @@ -681,37 +660,6 @@ typedef struct SStreamFillOperatorInfo { SStreamFillInfo* pFillInfo; } SStreamFillOperatorInfo; -typedef struct STimeSliceOperatorInfo { - SSDataBlock* pRes; - STimeWindow win; - SInterval interval; - int64_t current; - SArray* pPrevRow; // SArray - SArray* pNextRow; // SArray - SArray* pLinearInfo; // SArray - bool isPrevRowSet; - bool isNextRowSet; - int32_t fillType; // fill type - SColumn tsCol; // primary timestamp column - SExprSupp scalarSup; // scalar calculation - struct SFillColInfo* pFillColInfo; // fill column info -} STimeSliceOperatorInfo; - -typedef struct SStateWindowOperatorInfo { - // SOptrBasicInfo should be first, SAggSupporter should be second for stream encode - SOptrBasicInfo binfo; - SAggSupporter aggSup; - SExprSupp scalarSup; - - SGroupResInfo groupResInfo; - SWindowRowsSup winSup; - SColumn stateCol; // start row index - bool hasKey; - SStateKeys stateKey; - int32_t tsSlotId; // primary timestamp column slot id - STimeWindowAggSupp twAggSup; -} SStateWindowOperatorInfo; - #define OPTR_IS_OPENED(_optr) (((_optr)->status & OP_OPENED) == OP_OPENED) #define OPTR_SET_OPENED(_optr) ((_optr)->status |= OP_OPENED) @@ -726,6 +674,7 @@ void cleanupBasicInfo(SOptrBasicInfo* pInfo); int32_t initExprSupp(SExprSupp* pSup, SExprInfo* pExprInfo, int32_t numOfExpr); void cleanupExprSupp(SExprSupp* pSup); void destroyExprInfo(SExprInfo* pExpr, int32_t numOfExprs); + int32_t initAggInfo(SExprSupp* pSup, SAggSupporter* pAggSup, SExprInfo* pExprInfo, int32_t numOfCols, size_t keyBufSize, const char* pkey); void initResultSizeInfo(SResultInfo* pResultInfo, int32_t numOfRows); @@ -735,12 +684,12 @@ void doBuildStreamResBlock(SOperatorInfo* pOperator, SOptrBasicInfo* pbInfo, SGr void doBuildResultDatablock(SOperatorInfo* pOperator, SOptrBasicInfo* pbInfo, SGroupResInfo* pGroupResInfo, SDiskbasedBuf* pBuf); -bool hasLimitOffsetInfo(SLimitInfo* pLimitInfo); -void initLimitInfo(const SNode* pLimit, const SNode* pSLimit, SLimitInfo* pLimitInfo); +bool hasLimitOffsetInfo(SLimitInfo* pLimitInfo); +void initLimitInfo(const SNode* pLimit, const SNode* pSLimit, SLimitInfo* pLimitInfo); void applyLimitOffset(SLimitInfo* pLimitInfo, SSDataBlock* pBlock, SExecTaskInfo* pTaskInfo, SOperatorInfo* pOperator); -void doApplyFunctions(SExecTaskInfo* taskInfo, SqlFunctionCtx* pCtx, SColumnInfoData* pTimeWindowData, int32_t offset, - int32_t forwardStep, int32_t numOfTotal, int32_t numOfOutput); +void applyAggFunctionOnPartialTuples(SExecTaskInfo* taskInfo, SqlFunctionCtx* pCtx, SColumnInfoData* pTimeWindowData, + int32_t offset, int32_t forwardStep, int32_t numOfTotal, int32_t numOfOutput); int32_t extractDataBlockFromFetchRsp(SSDataBlock* pRes, char* pData, SArray* pColList, char** pNextStart); void updateLoadRemoteInfo(SLoadRemoteDataInfo* pInfo, int32_t numOfRows, int32_t dataLen, int64_t startTs, @@ -751,7 +700,7 @@ STimeWindow getFirstQualifiedTimeWindow(int64_t ts, STimeWindow* pWindow, SInter int32_t getTableScanInfo(SOperatorInfo* pOperator, int32_t* order, int32_t* scanFlag); int32_t getBufferPgSize(int32_t rowSize, uint32_t* defaultPgsz, uint32_t* defaultBufsz); -void doDestroyExchangeOperatorInfo(void* param); +extern void doDestroyExchangeOperatorInfo(void* param); void setOperatorCompleted(SOperatorInfo* pOperator); void setOperatorInfo(SOperatorInfo* pOperator, const char* name, int32_t type, bool blocking, int32_t status, @@ -764,79 +713,73 @@ void cleanupAggSup(SAggSupporter* pAggSup); void appendOneRowToDataBlock(SSDataBlock* pBlock, STupleHandle* pTupleHandle); void setTbNameColData(const SSDataBlock* pBlock, SColumnInfoData* pColInfoData, int32_t functionId, const char* name); -SSDataBlock* loadNextDataBlock(void* param); - void setResultRowInitCtx(SResultRow* pResult, SqlFunctionCtx* pCtx, int32_t numOfOutput, int32_t* rowEntryInfoOffset); SResultRow* doSetResultOutBufByKey(SDiskbasedBuf* pResultBuf, SResultRowInfo* pResultRowInfo, char* pData, int16_t bytes, bool masterscan, uint64_t groupId, SExecTaskInfo* pTaskInfo, bool isIntervalQuery, SAggSupporter* pSup); - +// operator creater functions +// clang-format off SOperatorInfo* createExchangeOperatorInfo(void* pTransporter, SExchangePhysiNode* pExNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createTableScanOperatorInfo(STableScanPhysiNode* pTableScanNode, SReadHandle* pHandle, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createTagScanOperatorInfo(SReadHandle* pReadHandle, STagScanPhysiNode* pPhyNode, - STableListInfo* pTableListInfo, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createSysTableScanOperatorInfo(void* readHandle, SSystemTableScanPhysiNode* pScanPhyNode, - const char* pUser, SExecTaskInfo* pTaskInfo); +SOperatorInfo* createTableScanOperatorInfo(STableScanPhysiNode* pTableScanNode, SReadHandle* pHandle, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createTagScanOperatorInfo(SReadHandle* pReadHandle, STagScanPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createSysTableScanOperatorInfo(void* readHandle, SSystemTableScanPhysiNode* pScanPhyNode, const char* pUser, SExecTaskInfo* pTaskInfo); SOperatorInfo* createAggregateOperatorInfo(SOperatorInfo* downstream, SAggPhysiNode* pNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createIndefinitOutputOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createProjectOperatorInfo(SOperatorInfo* downstream, SProjectPhysiNode* pProjPhyNode, - SExecTaskInfo* pTaskInfo); +SOperatorInfo* createIndefinitOutputOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createProjectOperatorInfo(SOperatorInfo* downstream, SProjectPhysiNode* pProjPhyNode, SExecTaskInfo* pTaskInfo); + SOperatorInfo* createSortOperatorInfo(SOperatorInfo* downstream, SSortPhysiNode* pSortNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createMultiwayMergeOperatorInfo(SOperatorInfo** dowStreams, size_t numStreams, - SMergePhysiNode* pMergePhysiNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createCacherowsScanOperator(SLastRowScanPhysiNode* pTableScanNode, SReadHandle* readHandle, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createIntervalOperatorInfo(SOperatorInfo* downstream, SIntervalPhysiNode* pPhyNode, - SExecTaskInfo* pTaskInfo, bool isStream); -SOperatorInfo* createMergeIntervalOperatorInfo(SOperatorInfo* downstream, SMergeIntervalPhysiNode* pIntervalPhyNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createMergeAlignedIntervalOperatorInfo(SOperatorInfo* downstream, SMergeAlignedIntervalPhysiNode* pNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStreamFinalIntervalOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, - SExecTaskInfo* pTaskInfo, int32_t numOfChild); -SOperatorInfo* createSessionAggOperatorInfo(SOperatorInfo* downstream, SSessionWinodwPhysiNode* pSessionNode, - SExecTaskInfo* pTaskInfo); +SOperatorInfo* createMultiwayMergeOperatorInfo(SOperatorInfo** dowStreams, size_t numStreams, SMergePhysiNode* pMergePhysiNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createCacherowsScanOperator(SLastRowScanPhysiNode* pTableScanNode, SReadHandle* readHandle, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createIntervalOperatorInfo(SOperatorInfo* downstream, SIntervalPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo, bool isStream); + +SOperatorInfo* createMergeIntervalOperatorInfo(SOperatorInfo* downstream, SMergeIntervalPhysiNode* pIntervalPhyNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createMergeAlignedIntervalOperatorInfo(SOperatorInfo* downstream, SMergeAlignedIntervalPhysiNode* pNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createStreamFinalIntervalOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo, int32_t numOfChild); + +SOperatorInfo* createSessionAggOperatorInfo(SOperatorInfo* downstream, SSessionWinodwPhysiNode* pSessionNode, SExecTaskInfo* pTaskInfo); + SOperatorInfo* createGroupOperatorInfo(SOperatorInfo* downstream, SAggPhysiNode* pAggNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createDataBlockInfoScanOperator(SReadHandle* readHandle, SBlockDistScanPhysiNode* pBlockScanNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhysiNode* pTableScanNode, SNode* pTagCond, - SExecTaskInfo* pTaskInfo); +SOperatorInfo* createDataBlockInfoScanOperator(SReadHandle* readHandle, SBlockDistScanPhysiNode* pBlockScanNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhysiNode* pTableScanNode, SNode* pTagCond, SExecTaskInfo* pTaskInfo); SOperatorInfo* createRawScanOperatorInfo(SReadHandle* pHandle, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createFillOperatorInfo(SOperatorInfo* downstream, SFillPhysiNode* pPhyFillNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStatewindowOperatorInfo(SOperatorInfo* downstream, SStateWinodwPhysiNode* pStateNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createPartitionOperatorInfo(SOperatorInfo* downstream, SPartitionPhysiNode* pPartNode, - SExecTaskInfo* pTaskInfo); +SOperatorInfo* createFillOperatorInfo(SOperatorInfo* downstream, SFillPhysiNode* pPhyFillNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStreamPartitionOperatorInfo(SOperatorInfo* downstream, SStreamPartitionPhysiNode* pPartNode, - SExecTaskInfo* pTaskInfo); +SOperatorInfo* createStatewindowOperatorInfo(SOperatorInfo* downstream, SStateWinodwPhysiNode* pStateNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createPartitionOperatorInfo(SOperatorInfo* downstream, SPartitionPhysiNode* pPartNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createStreamPartitionOperatorInfo(SOperatorInfo* downstream, SStreamPartitionPhysiNode* pPartNode, SExecTaskInfo* pTaskInfo); SOperatorInfo* createTimeSliceOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createMergeJoinOperatorInfo(SOperatorInfo** pDownstream, int32_t numOfDownstream, - SSortMergeJoinPhysiNode* pJoinNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStreamSessionAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStreamFinalSessionAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, - SExecTaskInfo* pTaskInfo, int32_t numOfChild); -SOperatorInfo* createStreamIntervalOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, - SExecTaskInfo* pTaskInfo); +SOperatorInfo* createMergeJoinOperatorInfo(SOperatorInfo** pDownstream, int32_t numOfDownstream, SSortMergeJoinPhysiNode* pJoinNode, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStreamStateAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFillPhysiNode* pPhyFillNode, - SExecTaskInfo* pTaskInfo); +SOperatorInfo* createStreamSessionAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createStreamFinalSessionAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo, int32_t numOfChild); + +SOperatorInfo* createStreamIntervalOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createStreamStateAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFillPhysiNode* pPhyFillNode, SExecTaskInfo* pTaskInfo); +// clang-format on int32_t projectApplyFunctions(SExprInfo* pExpr, SSDataBlock* pResult, SSDataBlock* pSrcBlock, SqlFunctionCtx* pCtx, int32_t numOfOutput, SArray* pPseudoList); diff --git a/source/libs/executor/src/cachescanoperator.c b/source/libs/executor/src/cachescanoperator.c index 873089023a..cdd744bded 100644 --- a/source/libs/executor/src/cachescanoperator.c +++ b/source/libs/executor/src/cachescanoperator.c @@ -59,7 +59,7 @@ SOperatorInfo* createCacherowsScanOperator(SLastRowScanPhysiNode* pScanNode, SRe pInfo->readHandle = *readHandle; SDataBlockDescNode* pDescNode = pScanNode->scan.node.pOutputDataBlockDesc; - pInfo->pRes = createResDataBlock(pDescNode); + pInfo->pRes = createDataBlockFromDescNode(pDescNode); int32_t numOfCols = 0; code = diff --git a/source/libs/executor/src/exchangeoperator.c b/source/libs/executor/src/exchangeoperator.c index d8c85c5ffb..280880c077 100644 --- a/source/libs/executor/src/exchangeoperator.c +++ b/source/libs/executor/src/exchangeoperator.c @@ -303,7 +303,7 @@ SOperatorInfo* createExchangeOperatorInfo(void* pTransporter, SExchangePhysiNode } tsem_init(&pInfo->ready, 0, 0); - pInfo->pDummyBlock = createResDataBlock(pExNode->node.pOutputDataBlockDesc); + pInfo->pDummyBlock = createDataBlockFromDescNode(pExNode->node.pOutputDataBlockDesc); pInfo->pResultBlockList = taosArrayInit(64, POINTER_BYTES); pInfo->pRecycledBlocks = taosArrayInit(64, POINTER_BYTES); diff --git a/source/libs/executor/src/executil.c b/source/libs/executor/src/executil.c index 65937ae1bc..08e6e4792b 100644 --- a/source/libs/executor/src/executil.c +++ b/source/libs/executor/src/executil.c @@ -208,7 +208,7 @@ SArray* createSortInfo(SNodeList* pNodeList) { return pList; } -SSDataBlock* createResDataBlock(SDataBlockDescNode* pNode) { +SSDataBlock* createDataBlockFromDescNode(SDataBlockDescNode* pNode) { int32_t numOfCols = LIST_LENGTH(pNode->pSlots); SSDataBlock* pBlock = createDataBlock(); diff --git a/source/libs/executor/src/executorimpl.c b/source/libs/executor/src/executorimpl.c index 7d8c7da78d..5eaa8ba8dd 100644 --- a/source/libs/executor/src/executorimpl.c +++ b/source/libs/executor/src/executorimpl.c @@ -76,6 +76,15 @@ static UNUSED_FUNC void* u_realloc(void* p, size_t __size) { #define CLEAR_QUERY_STATUS(q, st) ((q)->status &= (~(st))) #define QUERY_IS_INTERVAL_QUERY(_q) ((_q)->interval.interval > 0) +typedef struct SAggOperatorInfo { + SOptrBasicInfo binfo; + SAggSupporter aggSup; + STableQueryInfo* current; + uint64_t groupId; + SGroupResInfo groupResInfo; + SExprSupp scalarExprSup; +} SAggOperatorInfo; + int32_t getMaximumIdleDurationSec() { return tsShellActivityTimer * 2; } static void setBlockSMAInfo(SqlFunctionCtx* pCtx, SExprInfo* pExpr, SSDataBlock* pBlock); @@ -316,8 +325,8 @@ static void functionCtxRestore(SqlFunctionCtx* pCtx, SFunctionCtxStatus* pStatus pCtx->input.startRowIndex = pStatus->startOffset; } -void doApplyFunctions(SExecTaskInfo* taskInfo, SqlFunctionCtx* pCtx, SColumnInfoData* pTimeWindowData, int32_t offset, - int32_t forwardStep, int32_t numOfTotal, int32_t numOfOutput) { +void applyAggFunctionOnPartialTuples(SExecTaskInfo* taskInfo, SqlFunctionCtx* pCtx, SColumnInfoData* pTimeWindowData, + int32_t offset, int32_t forwardStep, int32_t numOfTotal, int32_t numOfOutput) { for (int32_t k = 0; k < numOfOutput; ++k) { // keep it temporarily SFunctionCtxStatus status = {0}; @@ -2039,7 +2048,7 @@ SOperatorInfo* createAggregateOperatorInfo(SOperatorInfo* downstream, SAggPhysiN goto _error; } - SSDataBlock* pResBlock = createResDataBlock(pAggNode->node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pAggNode->node.pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); size_t keyBufSize = sizeof(int64_t) + sizeof(int64_t) + POINTER_BYTES; @@ -2213,7 +2222,7 @@ SOperatorInfo* createFillOperatorInfo(SOperatorInfo* downstream, SFillPhysiNode* goto _error; } - pInfo->pRes = createResDataBlock(pPhyFillNode->node.pOutputDataBlockDesc); + pInfo->pRes = createDataBlockFromDescNode(pPhyFillNode->node.pOutputDataBlockDesc); SExprInfo* pExprInfo = createExprInfo(pPhyFillNode->pFillExprs, NULL, &pInfo->numOfExpr); pOperator->exprSupp.pExprInfo = pExprInfo; @@ -2512,7 +2521,7 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo return NULL; } - pOperator = createTagScanOperatorInfo(pHandle, pScanPhyNode, pTableListInfo, pTaskInfo); + pOperator = createTagScanOperatorInfo(pHandle, pScanPhyNode, pTaskInfo); } else if (QUERY_NODE_PHYSICAL_PLAN_BLOCK_DIST_SCAN == type) { SBlockDistScanPhysiNode* pBlockNode = (SBlockDistScanPhysiNode*)pPhyNode; diff --git a/source/libs/executor/src/groupoperator.c b/source/libs/executor/src/groupoperator.c index 8fd37c3b14..bbf9bd2a27 100644 --- a/source/libs/executor/src/groupoperator.c +++ b/source/libs/executor/src/groupoperator.c @@ -314,7 +314,7 @@ static void doHashGroupbyAgg(SOperatorInfo* pOperator, SSDataBlock* pBlock) { } int32_t rowIndex = j - num; - doApplyFunctions(pTaskInfo, pCtx, NULL, rowIndex, num, pBlock->info.rows, pOperator->exprSupp.numOfExprs); + applyAggFunctionOnPartialTuples(pTaskInfo, pCtx, NULL, rowIndex, num, pBlock->info.rows, pOperator->exprSupp.numOfExprs); // assign the group keys or user input constant values if required doAssignGroupKeys(pCtx, pOperator->exprSupp.numOfExprs, pBlock->info.rows, rowIndex); @@ -331,7 +331,7 @@ static void doHashGroupbyAgg(SOperatorInfo* pOperator, SSDataBlock* pBlock) { } int32_t rowIndex = pBlock->info.rows - num; - doApplyFunctions(pTaskInfo, pCtx, NULL, rowIndex, num, pBlock->info.rows, pOperator->exprSupp.numOfExprs); + applyAggFunctionOnPartialTuples(pTaskInfo, pCtx, NULL, rowIndex, num, pBlock->info.rows, pOperator->exprSupp.numOfExprs); doAssignGroupKeys(pCtx, pOperator->exprSupp.numOfExprs, pBlock->info.rows, rowIndex); } } @@ -431,7 +431,7 @@ SOperatorInfo* createGroupOperatorInfo(SOperatorInfo* downstream, SAggPhysiNode* goto _error; } - SSDataBlock* pResBlock = createResDataBlock(pAggNode->node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pAggNode->node.pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); int32_t numOfScalarExpr = 0; @@ -823,7 +823,7 @@ SOperatorInfo* createPartitionOperatorInfo(SOperatorInfo* downstream, SPartition uint32_t defaultPgsz = 0; uint32_t defaultBufsz = 0; - pInfo->binfo.pRes = createResDataBlock(pPartNode->node.pOutputDataBlockDesc); + pInfo->binfo.pRes = createDataBlockFromDescNode(pPartNode->node.pOutputDataBlockDesc); getBufferPgSize(pInfo->binfo.pRes->info.rowSize, &defaultPgsz, &defaultBufsz); if (!osTempSpaceAvailable()) { @@ -1119,7 +1119,7 @@ SOperatorInfo* createStreamPartitionOperatorInfo(SOperatorInfo* downstream, SStr } pInfo->partitionSup.needCalc = true; - pInfo->binfo.pRes = createResDataBlock(pPartNode->part.node.pOutputDataBlockDesc); + pInfo->binfo.pRes = createDataBlockFromDescNode(pPartNode->part.node.pOutputDataBlockDesc); if (pInfo->binfo.pRes == NULL) { goto _error; } diff --git a/source/libs/executor/src/joinoperator.c b/source/libs/executor/src/joinoperator.c index a1b44307d4..3839af9913 100644 --- a/source/libs/executor/src/joinoperator.c +++ b/source/libs/executor/src/joinoperator.c @@ -87,7 +87,7 @@ SOperatorInfo* createMergeJoinOperatorInfo(SOperatorInfo** pDownstream, int32_t } int32_t numOfCols = 0; - SSDataBlock* pResBlock = createResDataBlock(pJoinNode->node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pJoinNode->node.pOutputDataBlockDesc); SExprInfo* pExprInfo = createExprInfo(pJoinNode->pTargets, NULL, &numOfCols); initResultSizeInfo(&pOperator->resultInfo, 4096); diff --git a/source/libs/executor/src/projectoperator.c b/source/libs/executor/src/projectoperator.c index ada7964c67..819997c521 100644 --- a/source/libs/executor/src/projectoperator.c +++ b/source/libs/executor/src/projectoperator.c @@ -85,7 +85,7 @@ SOperatorInfo* createProjectOperatorInfo(SOperatorInfo* downstream, SProjectPhys int32_t numOfCols = 0; SExprInfo* pExprInfo = createExprInfo(pProjPhyNode->pProjections, NULL, &numOfCols); - SSDataBlock* pResBlock = createResDataBlock(pProjPhyNode->node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pProjPhyNode->node.pOutputDataBlockDesc); initLimitInfo(pProjPhyNode->node.pLimit, pProjPhyNode->node.pSlimit, &pInfo->limitInfo); pInfo->binfo.pRes = pResBlock; @@ -385,7 +385,7 @@ SOperatorInfo* createIndefinitOutputOperatorInfo(SOperatorInfo* downstream, SPhy } } - SSDataBlock* pResBlock = createResDataBlock(pPhyNode->node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pPhyNode->node.pOutputDataBlockDesc); int32_t numOfRows = 4096; size_t keyBufSize = sizeof(int64_t) + sizeof(int64_t) + POINTER_BYTES; diff --git a/source/libs/executor/src/scanoperator.c b/source/libs/executor/src/scanoperator.c index 7988c555e9..c0bea731bd 100644 --- a/source/libs/executor/src/scanoperator.c +++ b/source/libs/executor/src/scanoperator.c @@ -885,7 +885,7 @@ SOperatorInfo* createTableScanOperatorInfo(STableScanPhysiNode* pTableScanNode, pInfo->base.dataBlockLoadFlag = pTableScanNode->dataRequired; initResultSizeInfo(&pOperator->resultInfo, 4096); - pInfo->pResBlock = createResDataBlock(pDescNode); + pInfo->pResBlock = createDataBlockFromDescNode(pDescNode); blockDataEnsureCapacity(pInfo->pResBlock, pOperator->resultInfo.capacity); code = filterInitFromNode((SNode*)pTableScanNode->scan.node.pConditions, &pOperator->exprSupp.pFilterInfo, 0); @@ -2352,7 +2352,7 @@ SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhys goto _error; } - pInfo->pRes = createResDataBlock(pDescNode); + pInfo->pRes = createDataBlockFromDescNode(pDescNode); pInfo->pUpdateRes = createSpecialDataBlock(STREAM_CLEAR); pInfo->scanMode = STREAM_SCAN_FROM_READERHANDLE; pInfo->windowSup = (SWindowSupporter){.pStreamAggSup = NULL, .gap = -1, .parentType = QUERY_NODE_PHYSICAL_PLAN}; @@ -2476,8 +2476,7 @@ static void destroyTagScanOperatorInfo(void* param) { taosMemoryFreeClear(param); } -SOperatorInfo* createTagScanOperatorInfo(SReadHandle* pReadHandle, STagScanPhysiNode* pPhyNode, - STableListInfo* pTableListInfo, SExecTaskInfo* pTaskInfo) { +SOperatorInfo* createTagScanOperatorInfo(SReadHandle* pReadHandle, STagScanPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo) { STagScanInfo* pInfo = taosMemoryCalloc(1, sizeof(STagScanInfo)); SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo)); if (pInfo == NULL || pOperator == NULL) { @@ -2499,7 +2498,7 @@ SOperatorInfo* createTagScanOperatorInfo(SReadHandle* pReadHandle, STagScanPhysi goto _error; } - pInfo->pRes = createResDataBlock(pDescNode); + pInfo->pRes = createDataBlockFromDescNode(pDescNode); pInfo->readHandle = *pReadHandle; pInfo->curPos = 0; @@ -2613,7 +2612,7 @@ SArray* generateSortByTsInfo(SArray* colMatchInfo, int32_t order) { return pList; } -int32_t dumpSQueryTableCond(const SQueryTableDataCond* src, SQueryTableDataCond* dst) { +int32_t dumpQueryTableCond(const SQueryTableDataCond* src, SQueryTableDataCond* dst) { memcpy((void*)dst, (void*)src, sizeof(SQueryTableDataCond)); dst->colList = taosMemoryCalloc(src->numOfCols, sizeof(SColumnInfo)); for (int i = 0; i < src->numOfCols; i++) { @@ -2664,7 +2663,7 @@ int32_t startGroupTableMergeScan(SOperatorInfo* pOperator) { taosArrayPush(pInfo->sortSourceParams, ¶m); SQueryTableDataCond cond; - dumpSQueryTableCond(&pInfo->base.cond, &cond); + dumpQueryTableCond(&pInfo->base.cond, &cond); taosArrayPush(pInfo->queryConds, &cond); } @@ -2900,7 +2899,7 @@ SOperatorInfo* createTableMergeScanOperatorInfo(STableScanPhysiNode* pTableScanN } initResultSizeInfo(&pOperator->resultInfo, 1024); - pInfo->pResBlock = createResDataBlock(pDescNode); + pInfo->pResBlock = createDataBlockFromDescNode(pDescNode); blockDataEnsureCapacity(pInfo->pResBlock, pOperator->resultInfo.capacity); pInfo->sortSourceParams = taosArrayInit(64, sizeof(STableMergeScanSortSourceParam)); diff --git a/source/libs/executor/src/sortoperator.c b/source/libs/executor/src/sortoperator.c index 02cd0fe696..f2c8dc5083 100644 --- a/source/libs/executor/src/sortoperator.c +++ b/source/libs/executor/src/sortoperator.c @@ -47,7 +47,7 @@ SOperatorInfo* createSortOperatorInfo(SOperatorInfo* downstream, SSortPhysiNode* SDataBlockDescNode* pDescNode = pSortNode->node.pOutputDataBlockDesc; int32_t numOfCols = 0; - SSDataBlock* pResBlock = createResDataBlock(pDescNode); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pDescNode); SExprInfo* pExprInfo = createExprInfo(pSortNode->pExprs, NULL, &numOfCols); int32_t numOfOutputCols = 0; @@ -509,7 +509,7 @@ SOperatorInfo* createGroupSortOperatorInfo(SOperatorInfo* downstream, SGroupSort initResultSizeInfo(&pOperator->resultInfo, 1024); pOperator->exprSupp.pCtx = createSqlFunctionCtx(pExprInfo, numOfCols, &pOperator->exprSupp.rowEntryInfoOffset); - pInfo->binfo.pRes = createResDataBlock(pDescNode); + pInfo->binfo.pRes = createDataBlockFromDescNode(pDescNode); blockDataEnsureCapacity(pInfo->binfo.pRes, pOperator->resultInfo.capacity); int32_t numOfOutputCols = 0; @@ -766,7 +766,7 @@ SOperatorInfo* createMultiwayMergeOperatorInfo(SOperatorInfo** downStreams, size } initLimitInfo(pMergePhyNode->node.pLimit, pMergePhyNode->node.pSlimit, &pInfo->limitInfo); - pInfo->binfo.pRes = createResDataBlock(pDescNode); + pInfo->binfo.pRes = createDataBlockFromDescNode(pDescNode); int32_t rowSize = pInfo->binfo.pRes->info.rowSize; ASSERT(rowSize < 100 * 1024 * 1024); @@ -779,7 +779,7 @@ SOperatorInfo* createMultiwayMergeOperatorInfo(SOperatorInfo** downStreams, size } SPhysiNode* pChildNode = (SPhysiNode*)nodesListGetNode(pPhyNode->pChildren, 0); - SSDataBlock* pInputBlock = createResDataBlock(pChildNode->pOutputDataBlockDesc); + SSDataBlock* pInputBlock = createDataBlockFromDescNode(pChildNode->pOutputDataBlockDesc); initResultSizeInfo(&pOperator->resultInfo, 4096); blockDataEnsureCapacity(pInfo->binfo.pRes, pOperator->resultInfo.capacity); diff --git a/source/libs/executor/src/sysscanoperator.c b/source/libs/executor/src/sysscanoperator.c index eea2549a42..7ef2668804 100644 --- a/source/libs/executor/src/sysscanoperator.c +++ b/source/libs/executor/src/sysscanoperator.c @@ -1411,7 +1411,7 @@ SOperatorInfo* createSysTableScanOperatorInfo(void* readHandle, SSystemTableScan pInfo->pUser = taosMemoryStrDup((void*)pUser); pInfo->sysInfo = pScanPhyNode->sysInfo; pInfo->showRewrite = pScanPhyNode->showRewrite; - pInfo->pRes = createResDataBlock(pDescNode); + pInfo->pRes = createDataBlockFromDescNode(pDescNode); pInfo->pCondition = pScanNode->node.pConditions; code = filterInitFromNode(pScanNode->node.pConditions, &pOperator->exprSupp.pFilterInfo, 0); @@ -1928,7 +1928,7 @@ SOperatorInfo* createDataBlockInfoScanOperator(SReadHandle* readHandle, SBlockDi pInfo->readHandle = *readHandle; pInfo->uid = pBlockScanNode->suid; - pInfo->pResBlock = createResDataBlock(pBlockScanNode->node.pOutputDataBlockDesc); + pInfo->pResBlock = createDataBlockFromDescNode(pBlockScanNode->node.pOutputDataBlockDesc); blockDataEnsureCapacity(pInfo->pResBlock, 1); int32_t numOfCols = 0; diff --git a/source/libs/executor/src/tfill.c b/source/libs/executor/src/tfill.c index 7674b9e479..ba826a23d2 100644 --- a/source/libs/executor/src/tfill.c +++ b/source/libs/executor/src/tfill.c @@ -1651,9 +1651,9 @@ SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFi } initResultSizeInfo(&pOperator->resultInfo, 4096); - pInfo->pRes = createResDataBlock(pPhyFillNode->node.pOutputDataBlockDesc); - pInfo->pSrcBlock = createResDataBlock(pPhyFillNode->node.pOutputDataBlockDesc); - pInfo->pPrevSrcBlock = createResDataBlock(pPhyFillNode->node.pOutputDataBlockDesc); + pInfo->pRes = createDataBlockFromDescNode(pPhyFillNode->node.pOutputDataBlockDesc); + pInfo->pSrcBlock = createDataBlockFromDescNode(pPhyFillNode->node.pOutputDataBlockDesc); + pInfo->pPrevSrcBlock = createDataBlockFromDescNode(pPhyFillNode->node.pOutputDataBlockDesc); blockDataEnsureCapacity(pInfo->pRes, pOperator->resultInfo.capacity); blockDataEnsureCapacity(pInfo->pSrcBlock, pOperator->resultInfo.capacity); blockDataEnsureCapacity(pInfo->pPrevSrcBlock, pOperator->resultInfo.capacity); diff --git a/source/libs/executor/src/timewindowoperator.c b/source/libs/executor/src/timewindowoperator.c index 013b8d39de..80c3c1c454 100644 --- a/source/libs/executor/src/timewindowoperator.c +++ b/source/libs/executor/src/timewindowoperator.c @@ -22,13 +22,37 @@ #include "tfill.h" #include "ttime.h" +#define IS_FINAL_OP(op) ((op)->isFinal) + +typedef struct SSessionAggOperatorInfo { + SOptrBasicInfo binfo; + SAggSupporter aggSup; + SGroupResInfo groupResInfo; + SWindowRowsSup winSup; + bool reptScan; // next round scan + int64_t gap; // session window gap + int32_t tsSlotId; // primary timestamp slot id + STimeWindowAggSupp twAggSup; +} SSessionAggOperatorInfo; + +typedef struct SStateWindowOperatorInfo { + SOptrBasicInfo binfo; + SAggSupporter aggSup; + SExprSupp scalarSup; + SGroupResInfo groupResInfo; + SWindowRowsSup winSup; + SColumn stateCol; // start row index + bool hasKey; + SStateKeys stateKey; + int32_t tsSlotId; // primary timestamp column slot id + STimeWindowAggSupp twAggSup; +} SStateWindowOperatorInfo; + typedef enum SResultTsInterpType { RESULT_ROW_START_INTERP = 1, RESULT_ROW_END_INTERP = 2, } SResultTsInterpType; -#define IS_FINAL_OP(op) ((op)->isFinal) - typedef struct SPullWindowInfo { STimeWindow window; uint64_t groupId; @@ -640,7 +664,7 @@ static void doInterpUnclosedTimeWindow(SOperatorInfo* pOperatorInfo, int32_t num setNotInterpoWindowKey(pSup->pCtx, numOfExprs, RESULT_ROW_START_INTERP); updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &w, true); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, 0, pBlock->info.rows, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, 0, pBlock->info.rows, numOfExprs); if (isResultRowInterpolated(pResult, RESULT_ROW_END_INTERP)) { @@ -937,7 +961,7 @@ static void hashIntervalAgg(SOperatorInfo* pOperatorInfo, SResultRowInfo* pResul } updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &win, true); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, forwardRows, pBlock->info.rows, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, forwardRows, pBlock->info.rows, numOfOutput); doCloseWindow(pResultRowInfo, pInfo, pResult); @@ -972,7 +996,7 @@ static void hashIntervalAgg(SOperatorInfo* pOperatorInfo, SResultRowInfo* pResul } #endif updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &nextWin, true); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, forwardRows, pBlock->info.rows, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, forwardRows, pBlock->info.rows, numOfOutput); doCloseWindow(pResultRowInfo, pInfo, pResult); } @@ -1140,7 +1164,7 @@ static void doStateWindowAggImpl(SOperatorInfo* pOperator, SStateWindowOperatorI } updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &window, false); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, pRowSup->numOfRows, pBlock->info.rows, numOfOutput); // here we start a new session window @@ -1165,7 +1189,7 @@ static void doStateWindowAggImpl(SOperatorInfo* pOperator, SStateWindowOperatorI } updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &pRowSup->win, false); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, pRowSup->numOfRows, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, pRowSup->numOfRows, pBlock->info.rows, numOfOutput); } @@ -1706,7 +1730,7 @@ SOperatorInfo* createIntervalOperatorInfo(SOperatorInfo* downstream, SIntervalPh goto _error; } - SSDataBlock* pResBlock = createResDataBlock(pPhyNode->window.node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pPhyNode->window.node.pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); SExprSupp* pSup = &pOperator->exprSupp; @@ -1845,7 +1869,7 @@ static void doSessionWindowAggImpl(SOperatorInfo* pOperator, SSessionAggOperator // pInfo->numOfRows data belong to the current session window updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &window, false); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, pRowSup->numOfRows, pBlock->info.rows, numOfOutput); // here we start a new session window @@ -1863,7 +1887,7 @@ static void doSessionWindowAggImpl(SOperatorInfo* pOperator, SSessionAggOperator } updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &pRowSup->win, false); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, pRowSup->numOfRows, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, pRowSup->startRowIndex, pRowSup->numOfRows, pBlock->info.rows, numOfOutput); } @@ -1938,552 +1962,6 @@ static SSDataBlock* doSessionWindowAgg(SOperatorInfo* pOperator) { return (pBInfo->pRes->info.rows == 0) ? NULL : pBInfo->pRes; } -static void doKeepPrevRows(STimeSliceOperatorInfo* pSliceInfo, const SSDataBlock* pBlock, int32_t rowIndex) { - int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); - for (int32_t i = 0; i < numOfCols; ++i) { - SColumnInfoData* pColInfoData = taosArrayGet(pBlock->pDataBlock, i); - - SGroupKeys* pkey = taosArrayGet(pSliceInfo->pPrevRow, i); - if (!colDataIsNull_s(pColInfoData, rowIndex)) { - pkey->isNull = false; - char* val = colDataGetData(pColInfoData, rowIndex); - if (!IS_VAR_DATA_TYPE(pkey->type)) { - memcpy(pkey->pData, val, pkey->bytes); - } else { - memcpy(pkey->pData, val, varDataLen(val)); - } - } else { - pkey->isNull = true; - } - } - - pSliceInfo->isPrevRowSet = true; -} - -static void doKeepNextRows(STimeSliceOperatorInfo* pSliceInfo, const SSDataBlock* pBlock, int32_t rowIndex) { - int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); - for (int32_t i = 0; i < numOfCols; ++i) { - SColumnInfoData* pColInfoData = taosArrayGet(pBlock->pDataBlock, i); - - SGroupKeys* pkey = taosArrayGet(pSliceInfo->pNextRow, i); - if (!colDataIsNull_s(pColInfoData, rowIndex)) { - pkey->isNull = false; - char* val = colDataGetData(pColInfoData, rowIndex); - if (!IS_VAR_DATA_TYPE(pkey->type)) { - memcpy(pkey->pData, val, pkey->bytes); - } else { - memcpy(pkey->pData, val, varDataLen(val)); - } - } else { - pkey->isNull = true; - } - } - - pSliceInfo->isNextRowSet = true; -} - -static void doKeepLinearInfo(STimeSliceOperatorInfo* pSliceInfo, const SSDataBlock* pBlock, int32_t rowIndex) { - int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); - for (int32_t i = 0; i < numOfCols; ++i) { - SColumnInfoData* pColInfoData = taosArrayGet(pBlock->pDataBlock, i); - SColumnInfoData* pTsCol = taosArrayGet(pBlock->pDataBlock, pSliceInfo->tsCol.slotId); - SFillLinearInfo* pLinearInfo = taosArrayGet(pSliceInfo->pLinearInfo, i); - - // null value is represented by using key = INT64_MIN for now. - // TODO: optimize to ignore null values for linear interpolation. - if (!pLinearInfo->isStartSet) { - if (!colDataIsNull_s(pColInfoData, rowIndex)) { - pLinearInfo->start.key = *(int64_t*)colDataGetData(pTsCol, rowIndex); - memcpy(pLinearInfo->start.val, colDataGetData(pColInfoData, rowIndex), pLinearInfo->bytes); - } - pLinearInfo->isStartSet = true; - } else if (!pLinearInfo->isEndSet) { - if (!colDataIsNull_s(pColInfoData, rowIndex)) { - pLinearInfo->end.key = *(int64_t*)colDataGetData(pTsCol, rowIndex); - memcpy(pLinearInfo->end.val, colDataGetData(pColInfoData, rowIndex), pLinearInfo->bytes); - } - pLinearInfo->isEndSet = true; - } else { - pLinearInfo->start.key = pLinearInfo->end.key; - memcpy(pLinearInfo->start.val, pLinearInfo->end.val, pLinearInfo->bytes); - - if (!colDataIsNull_s(pColInfoData, rowIndex)) { - pLinearInfo->end.key = *(int64_t*)colDataGetData(pTsCol, rowIndex); - memcpy(pLinearInfo->end.val, colDataGetData(pColInfoData, rowIndex), pLinearInfo->bytes); - } else { - pLinearInfo->end.key = INT64_MIN; - } - } - } - -} - -static bool genInterpolationResult(STimeSliceOperatorInfo* pSliceInfo, SExprSupp* pExprSup, SSDataBlock* pResBlock, bool beforeTs) { - int32_t rows = pResBlock->info.rows; - blockDataEnsureCapacity(pResBlock, rows + 1); - // todo set the correct primary timestamp column - - // output the result - bool hasInterp = true; - for (int32_t j = 0; j < pExprSup->numOfExprs; ++j) { - SExprInfo* pExprInfo = &pExprSup->pExprInfo[j]; - - int32_t dstSlot = pExprInfo->base.resSchema.slotId; - SColumnInfoData* pDst = taosArrayGet(pResBlock->pDataBlock, dstSlot); - - if (IS_TIMESTAMP_TYPE(pExprInfo->base.resSchema.type)) { - colDataAppend(pDst, rows, (char*)&pSliceInfo->current, false); - continue; - } - - int32_t srcSlot = pExprInfo->base.pParam[0].pCol->slotId; - switch (pSliceInfo->fillType) { - case TSDB_FILL_NULL: { - colDataAppendNULL(pDst, rows); - break; - } - - case TSDB_FILL_SET_VALUE: { - SVariant* pVar = &pSliceInfo->pFillColInfo[j].fillVal; - - if (pDst->info.type == TSDB_DATA_TYPE_FLOAT) { - float v = 0; - GET_TYPED_DATA(v, float, pVar->nType, &pVar->i); - colDataAppend(pDst, rows, (char*)&v, false); - } else if (pDst->info.type == TSDB_DATA_TYPE_DOUBLE) { - double v = 0; - GET_TYPED_DATA(v, double, pVar->nType, &pVar->i); - colDataAppend(pDst, rows, (char*)&v, false); - } else if (IS_SIGNED_NUMERIC_TYPE(pDst->info.type)) { - int64_t v = 0; - GET_TYPED_DATA(v, int64_t, pVar->nType, &pVar->i); - colDataAppend(pDst, rows, (char*)&v, false); - } - break; - } - - case TSDB_FILL_LINEAR: { - SFillLinearInfo* pLinearInfo = taosArrayGet(pSliceInfo->pLinearInfo, srcSlot); - - SPoint start = pLinearInfo->start; - SPoint end = pLinearInfo->end; - SPoint current = {.key = pSliceInfo->current}; - - // do not interpolate before ts range, only increate pSliceInfo->current - if (beforeTs && !pLinearInfo->isEndSet) { - return true; - } - - if (!pLinearInfo->isStartSet || !pLinearInfo->isEndSet) { - hasInterp = false; - break; - } - - if (start.key == INT64_MIN || end.key == INT64_MIN) { - colDataAppendNULL(pDst, rows); - break; - } - - current.val = taosMemoryCalloc(pLinearInfo->bytes, 1); - taosGetLinearInterpolationVal(¤t, pLinearInfo->type, &start, &end, pLinearInfo->type); - colDataAppend(pDst, rows, (char*)current.val, false); - - taosMemoryFree(current.val); - break; - } - case TSDB_FILL_PREV: { - if (!pSliceInfo->isPrevRowSet) { - hasInterp = false; - break; - } - - SGroupKeys* pkey = taosArrayGet(pSliceInfo->pPrevRow, srcSlot); - if (pkey->isNull == false) { - colDataAppend(pDst, rows, pkey->pData, false); - } else { - colDataAppendNULL(pDst, rows); - } - break; - } - - case TSDB_FILL_NEXT: { - if (!pSliceInfo->isNextRowSet) { - hasInterp = false; - break; - } - - SGroupKeys* pkey = taosArrayGet(pSliceInfo->pNextRow, srcSlot); - if (pkey->isNull == false) { - colDataAppend(pDst, rows, pkey->pData, false); - } else { - colDataAppendNULL(pDst, rows); - } - break; - } - - case TSDB_FILL_NONE: - default: - break; - } - } - - if (hasInterp) { - pResBlock->info.rows += 1; - } - - return hasInterp; -} - -static void addCurrentRowToResult(STimeSliceOperatorInfo* pSliceInfo, SExprSupp* pExprSup, SSDataBlock* pResBlock, - SSDataBlock* pSrcBlock, int32_t index) { - blockDataEnsureCapacity(pResBlock, pResBlock->info.rows + 1); - for (int32_t j = 0; j < pExprSup->numOfExprs; ++j) { - SExprInfo* pExprInfo = &pExprSup->pExprInfo[j]; - - int32_t dstSlot = pExprInfo->base.resSchema.slotId; - SColumnInfoData* pDst = taosArrayGet(pResBlock->pDataBlock, dstSlot); - - if (IS_TIMESTAMP_TYPE(pExprInfo->base.resSchema.type)) { - colDataAppend(pDst, pResBlock->info.rows, (char*)&pSliceInfo->current, false); - } else { - int32_t srcSlot = pExprInfo->base.pParam[0].pCol->slotId; - SColumnInfoData* pSrc = taosArrayGet(pSrcBlock->pDataBlock, srcSlot); - - if (colDataIsNull_s(pSrc, index)) { - colDataAppendNULL(pDst, pResBlock->info.rows); - continue; - } - - char* v = colDataGetData(pSrc, index); - colDataAppend(pDst, pResBlock->info.rows, v, false); - } - } - - pResBlock->info.rows += 1; - return; -} - - -static int32_t initPrevRowsKeeper(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { - if (pInfo->pPrevRow != NULL) { - return TSDB_CODE_SUCCESS; - } - - pInfo->pPrevRow = taosArrayInit(4, sizeof(SGroupKeys)); - if (pInfo->pPrevRow == NULL) { - return TSDB_CODE_OUT_OF_MEMORY; - } - - int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); - for (int32_t i = 0; i < numOfCols; ++i) { - SColumnInfoData* pColInfo = taosArrayGet(pBlock->pDataBlock, i); - - SGroupKeys key = {0}; - key.bytes = pColInfo->info.bytes; - key.type = pColInfo->info.type; - key.isNull = false; - key.pData = taosMemoryCalloc(1, pColInfo->info.bytes); - taosArrayPush(pInfo->pPrevRow, &key); - } - - pInfo->isPrevRowSet = false; - - return TSDB_CODE_SUCCESS; -} - -static int32_t initNextRowsKeeper(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { - if (pInfo->pNextRow != NULL) { - return TSDB_CODE_SUCCESS; - } - - pInfo->pNextRow = taosArrayInit(4, sizeof(SGroupKeys)); - if (pInfo->pNextRow == NULL) { - return TSDB_CODE_OUT_OF_MEMORY; - } - - int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); - for (int32_t i = 0; i < numOfCols; ++i) { - SColumnInfoData* pColInfo = taosArrayGet(pBlock->pDataBlock, i); - - SGroupKeys key = {0}; - key.bytes = pColInfo->info.bytes; - key.type = pColInfo->info.type; - key.isNull = false; - key.pData = taosMemoryCalloc(1, pColInfo->info.bytes); - taosArrayPush(pInfo->pNextRow, &key); - } - - pInfo->isNextRowSet = false; - - return TSDB_CODE_SUCCESS; -} - -static int32_t initFillLinearInfo(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { - if (pInfo->pLinearInfo != NULL) { - return TSDB_CODE_SUCCESS; - } - - pInfo->pLinearInfo = taosArrayInit(4, sizeof(SFillLinearInfo)); - if (pInfo->pLinearInfo == NULL) { - return TSDB_CODE_OUT_OF_MEMORY; - } - - int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); - for (int32_t i = 0; i < numOfCols; ++i) { - SColumnInfoData* pColInfo = taosArrayGet(pBlock->pDataBlock, i); - - SFillLinearInfo linearInfo = {0}; - linearInfo.start.key = INT64_MIN; - linearInfo.end.key = INT64_MIN; - linearInfo.start.val = taosMemoryCalloc(1, pColInfo->info.bytes); - linearInfo.end.val = taosMemoryCalloc(1, pColInfo->info.bytes); - linearInfo.isStartSet = false; - linearInfo.isEndSet = false; - linearInfo.type = pColInfo->info.type; - linearInfo.bytes = pColInfo->info.bytes; - taosArrayPush(pInfo->pLinearInfo, &linearInfo); - } - - return TSDB_CODE_SUCCESS; -} - -static int32_t initKeeperInfo(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { - int32_t code; - code = initPrevRowsKeeper(pInfo, pBlock); - if (code != TSDB_CODE_SUCCESS) { - return TSDB_CODE_FAILED; - } - - code = initNextRowsKeeper(pInfo, pBlock); - if (code != TSDB_CODE_SUCCESS) { - return TSDB_CODE_FAILED; - } - - code = initFillLinearInfo(pInfo, pBlock); - if (code != TSDB_CODE_SUCCESS) { - return TSDB_CODE_FAILED; - } - - return TSDB_CODE_SUCCESS; -} - -static SSDataBlock* doTimeslice(SOperatorInfo* pOperator) { - if (pOperator->status == OP_EXEC_DONE) { - return NULL; - } - - SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo; - - STimeSliceOperatorInfo* pSliceInfo = pOperator->info; - SSDataBlock* pResBlock = pSliceInfo->pRes; - SExprSupp* pSup = &pOperator->exprSupp; - - int32_t order = TSDB_ORDER_ASC; - SInterval* pInterval = &pSliceInfo->interval; - SOperatorInfo* downstream = pOperator->pDownstream[0]; - - blockDataCleanup(pResBlock); - - while (1) { - SSDataBlock* pBlock = downstream->fpSet.getNextFn(downstream); - if (pBlock == NULL) { - break; - } - - int32_t code = initKeeperInfo(pSliceInfo, pBlock); - if (code != TSDB_CODE_SUCCESS) { - T_LONG_JMP(pTaskInfo->env, code); - } - - // the pDataBlock are always the same one, no need to call this again - setInputDataBlock(pSup, pBlock, order, MAIN_SCAN, true); - - SColumnInfoData* pTsCol = taosArrayGet(pBlock->pDataBlock, pSliceInfo->tsCol.slotId); - for (int32_t i = 0; i < pBlock->info.rows; ++i) { - int64_t ts = *(int64_t*)colDataGetData(pTsCol, i); - - if (pSliceInfo->current > pSliceInfo->win.ekey) { - setOperatorCompleted(pOperator); - break; - } - - if (ts == pSliceInfo->current) { - addCurrentRowToResult(pSliceInfo, &pOperator->exprSupp, pResBlock, pBlock, i); - - doKeepPrevRows(pSliceInfo, pBlock, i); - doKeepLinearInfo(pSliceInfo, pBlock, i); - - pSliceInfo->current = - taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); - if (pSliceInfo->current > pSliceInfo->win.ekey) { - setOperatorCompleted(pOperator); - break; - } - } else if (ts < pSliceInfo->current) { - // in case of interpolation window starts and ends between two datapoints, fill(prev) need to interpolate - doKeepPrevRows(pSliceInfo, pBlock, i); - doKeepLinearInfo(pSliceInfo, pBlock, i); - - if (i < pBlock->info.rows - 1) { - // in case of interpolation window starts and ends between two datapoints, fill(next) need to interpolate - doKeepNextRows(pSliceInfo, pBlock, i + 1); - int64_t nextTs = *(int64_t*)colDataGetData(pTsCol, i + 1); - if (nextTs > pSliceInfo->current) { - while (pSliceInfo->current < nextTs && pSliceInfo->current <= pSliceInfo->win.ekey) { - if (!genInterpolationResult(pSliceInfo, &pOperator->exprSupp, pResBlock, false) && pSliceInfo->fillType == TSDB_FILL_LINEAR) { - break; - } else { - pSliceInfo->current = - taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); - } - } - - if (pSliceInfo->current > pSliceInfo->win.ekey) { - setOperatorCompleted(pOperator); - break; - } - } else { - // ignore current row, and do nothing - } - } else { // it is the last row of current block - doKeepPrevRows(pSliceInfo, pBlock, i); - } - } else { // ts > pSliceInfo->current - // in case of interpolation window starts and ends between two datapoints, fill(next) need to interpolate - doKeepNextRows(pSliceInfo, pBlock, i); - doKeepLinearInfo(pSliceInfo, pBlock, i); - - while (pSliceInfo->current < ts && pSliceInfo->current <= pSliceInfo->win.ekey) { - if (!genInterpolationResult(pSliceInfo, &pOperator->exprSupp, pResBlock, true) && pSliceInfo->fillType == TSDB_FILL_LINEAR) { - break; - } else { - pSliceInfo->current = - taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); - } - } - - // add current row if timestamp match - if (ts == pSliceInfo->current && pSliceInfo->current <= pSliceInfo->win.ekey) { - addCurrentRowToResult(pSliceInfo, &pOperator->exprSupp, pResBlock, pBlock, i); - doKeepPrevRows(pSliceInfo, pBlock, i); - - pSliceInfo->current = - taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); - } - - if (pSliceInfo->current > pSliceInfo->win.ekey) { - setOperatorCompleted(pOperator); - break; - } - } - } - } - - // check if need to interpolate after last datablock - // except for fill(next), fill(linear) - while (pSliceInfo->current <= pSliceInfo->win.ekey && pSliceInfo->fillType != TSDB_FILL_NEXT && - pSliceInfo->fillType != TSDB_FILL_LINEAR) { - genInterpolationResult(pSliceInfo, &pOperator->exprSupp, pResBlock, false); - pSliceInfo->current = - taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); - } - - // restore the value - setTaskStatus(pOperator->pTaskInfo, TASK_COMPLETED); - if (pResBlock->info.rows == 0) { - pOperator->status = OP_EXEC_DONE; - } - - return pResBlock->info.rows == 0 ? NULL : pResBlock; -} - -void destroyTimeSliceOperatorInfo(void* param) { - STimeSliceOperatorInfo* pInfo = (STimeSliceOperatorInfo*)param; - - pInfo->pRes = blockDataDestroy(pInfo->pRes); - - for (int32_t i = 0; i < taosArrayGetSize(pInfo->pPrevRow); ++i) { - SGroupKeys* pKey = taosArrayGet(pInfo->pPrevRow, i); - taosMemoryFree(pKey->pData); - } - taosArrayDestroy(pInfo->pPrevRow); - - for (int32_t i = 0; i < taosArrayGetSize(pInfo->pNextRow); ++i) { - SGroupKeys* pKey = taosArrayGet(pInfo->pNextRow, i); - taosMemoryFree(pKey->pData); - } - taosArrayDestroy(pInfo->pNextRow); - - for (int32_t i = 0; i < taosArrayGetSize(pInfo->pLinearInfo); ++i) { - SFillLinearInfo* pKey = taosArrayGet(pInfo->pLinearInfo, i); - taosMemoryFree(pKey->start.val); - taosMemoryFree(pKey->end.val); - } - taosArrayDestroy(pInfo->pLinearInfo); - - taosMemoryFree(pInfo->pFillColInfo); - taosMemoryFreeClear(param); -} - -SOperatorInfo* createTimeSliceOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo) { - STimeSliceOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(STimeSliceOperatorInfo)); - SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo)); - if (pOperator == NULL || pInfo == NULL) { - goto _error; - } - - SInterpFuncPhysiNode* pInterpPhyNode = (SInterpFuncPhysiNode*)pPhyNode; - SExprSupp* pSup = &pOperator->exprSupp; - - int32_t numOfExprs = 0; - SExprInfo* pExprInfo = createExprInfo(pInterpPhyNode->pFuncs, NULL, &numOfExprs); - int32_t code = initExprSupp(pSup, pExprInfo, numOfExprs); - if (code != TSDB_CODE_SUCCESS) { - goto _error; - } - - if (pInterpPhyNode->pExprs != NULL) { - int32_t num = 0; - SExprInfo* pScalarExprInfo = createExprInfo(pInterpPhyNode->pExprs, NULL, &num); - code = initExprSupp(&pInfo->scalarSup, pScalarExprInfo, num); - if (code != TSDB_CODE_SUCCESS) { - goto _error; - } - } - - pInfo->tsCol = extractColumnFromColumnNode((SColumnNode*)pInterpPhyNode->pTimeSeries); - pInfo->fillType = convertFillType(pInterpPhyNode->fillMode); - initResultSizeInfo(&pOperator->resultInfo, 4096); - - pInfo->pFillColInfo = createFillColInfo(pExprInfo, numOfExprs, NULL, 0, (SNodeListNode*)pInterpPhyNode->pFillValues); - pInfo->pLinearInfo = NULL; - pInfo->pRes = createResDataBlock(pPhyNode->pOutputDataBlockDesc); - pInfo->win = pInterpPhyNode->timeRange; - pInfo->interval.interval = pInterpPhyNode->interval; - pInfo->current = pInfo->win.skey; - - if (downstream->operatorType == QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN) { - STableScanInfo* pScanInfo = (STableScanInfo*)downstream->info; - pScanInfo->base.cond.twindows = pInfo->win; - pScanInfo->base.cond.type = TIMEWINDOW_RANGE_EXTERNAL; - } - - setOperatorInfo(pOperator, "TimeSliceOperator", QUERY_NODE_PHYSICAL_PLAN_INTERP_FUNC, false, OP_NOT_OPENED, pInfo, - pTaskInfo); - pOperator->fpSet = createOperatorFpSet(operatorDummyOpenFn, doTimeslice, NULL, destroyTimeSliceOperatorInfo, NULL); - - blockDataEnsureCapacity(pInfo->pRes, pOperator->resultInfo.capacity); - - code = appendDownstream(pOperator, &downstream, 1); - return pOperator; - -_error: - taosMemoryFree(pInfo); - taosMemoryFree(pOperator); - pTaskInfo->code = TSDB_CODE_OUT_OF_MEMORY; - return NULL; -} - SOperatorInfo* createStatewindowOperatorInfo(SOperatorInfo* downstream, SStateWinodwPhysiNode* pStateNode, SExecTaskInfo* pTaskInfo) { SStateWindowOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(SStateWindowOperatorInfo)); @@ -2528,7 +2006,7 @@ SOperatorInfo* createStatewindowOperatorInfo(SOperatorInfo* downstream, SStateWi goto _error; } - SSDataBlock* pResBlock = createResDataBlock(pStateNode->window.node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pStateNode->window.node.pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); initResultRowInfo(&pInfo->binfo.resultRowInfo); @@ -2588,7 +2066,7 @@ SOperatorInfo* createSessionAggOperatorInfo(SOperatorInfo* downstream, SSessionW int32_t numOfCols = 0; SExprInfo* pExprInfo = createExprInfo(pSessionNode->window.pFuncs, NULL, &numOfCols); - SSDataBlock* pResBlock = createResDataBlock(pSessionNode->window.node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pSessionNode->window.node.pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); int32_t code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); @@ -2975,7 +2453,7 @@ static void doStreamIntervalAggImpl(SOperatorInfo* pOperatorInfo, SSDataBlock* p tSimpleHashPut(pInfo->aggSup.pResultRowHashTable, &key, sizeof(SWinKey), NULL, 0); } updateTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &nextWin, true); - doApplyFunctions(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, forwardRows, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &pInfo->twAggSup.timeWindowData, startPos, forwardRows, pSDataBlock->info.rows, numOfOutput); SWinKey key = { .ts = nextWin.skey, @@ -3226,7 +2704,7 @@ SOperatorInfo* createStreamFinalIntervalOperatorInfo(SOperatorInfo* downstream, int32_t numOfCols = 0; SExprInfo* pExprInfo = createExprInfo(pIntervalPhyNode->window.pFuncs, NULL, &numOfCols); - SSDataBlock* pResBlock = createResDataBlock(pPhyNode->pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pPhyNode->pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); int32_t code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); @@ -3551,7 +3029,7 @@ static int32_t doOneWindowAggImpl(SColumnInfoData* pTimeWindowData, SResultWindo return TSDB_CODE_QRY_OUT_OF_MEMORY; } updateTimeWindowInfo(pTimeWindowData, &pCurWin->sessionWin.win, false); - doApplyFunctions(pTaskInfo, pSup->pCtx, pTimeWindowData, startIndex, winRows, rows, numOutput); + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, pTimeWindowData, startIndex, winRows, rows, numOutput); return TSDB_CODE_SUCCESS; } @@ -4066,7 +3544,7 @@ SOperatorInfo* createStreamSessionAggOperatorInfo(SOperatorInfo* downstream, SPh SExprSupp* pSup = &pOperator->exprSupp; SExprInfo* pExprInfo = createExprInfo(pSessionNode->window.pFuncs, NULL, &numOfCols); - SSDataBlock* pResBlock = createResDataBlock(pPhyNode->pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pPhyNode->pOutputDataBlockDesc); code = initBasicInfoEx(&pInfo->binfo, pSup, pExprInfo, numOfCols, pResBlock); if (code != TSDB_CODE_SUCCESS) { goto _error; @@ -4586,7 +4064,7 @@ SOperatorInfo* createStreamStateAggOperatorInfo(SOperatorInfo* downstream, SPhys SExprSupp* pSup = &pOperator->exprSupp; int32_t numOfCols = 0; SExprInfo* pExprInfo = createExprInfo(pStateNode->window.pFuncs, NULL, &numOfCols); - SSDataBlock* pResBlock = createResDataBlock(pPhyNode->pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pPhyNode->pOutputDataBlockDesc); code = initBasicInfoEx(&pInfo->binfo, pSup, pExprInfo, numOfCols, pResBlock); if (code != TSDB_CODE_SUCCESS) { goto _error; @@ -4695,7 +4173,7 @@ static void doMergeAlignedIntervalAggImpl(SOperatorInfo* pOperatorInfo, SResultR } updateTimeWindowInfo(&iaInfo->twAggSup.timeWindowData, &currWin, true); - doApplyFunctions(pTaskInfo, pSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, currPos - startPos, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, currPos - startPos, pBlock->info.rows, pSup->numOfExprs); finalizeResultRows(iaInfo->aggSup.pResultBuf, &pResultRowInfo->cur, pSup, pResultBlock, pTaskInfo); @@ -4715,7 +4193,7 @@ static void doMergeAlignedIntervalAggImpl(SOperatorInfo* pOperatorInfo, SResultR } updateTimeWindowInfo(&iaInfo->twAggSup.timeWindowData, &currWin, true); - doApplyFunctions(pTaskInfo, pSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, currPos - startPos, + applyAggFunctionOnPartialTuples(pTaskInfo, pSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, currPos - startPos, pBlock->info.rows, pSup->numOfExprs); } @@ -4873,7 +4351,7 @@ SOperatorInfo* createMergeAlignedIntervalOperatorInfo(SOperatorInfo* downstream, goto _error; } - SSDataBlock* pResBlock = createResDataBlock(pNode->window.node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pNode->window.node.pOutputDataBlockDesc); initBasicInfo(&iaInfo->binfo, pResBlock); initExecTimeWindowInfo(&iaInfo->twAggSup.timeWindowData, &iaInfo->win); @@ -5022,7 +4500,7 @@ static void doMergeIntervalAggImpl(SOperatorInfo* pOperatorInfo, SResultRowInfo* } updateTimeWindowInfo(&iaInfo->twAggSup.timeWindowData, &win, true); - doApplyFunctions(pTaskInfo, pExprSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, forwardRows, + applyAggFunctionOnPartialTuples(pTaskInfo, pExprSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, forwardRows, pBlock->info.rows, numOfOutput); doCloseWindow(pResultRowInfo, iaInfo, pResult); @@ -5054,7 +4532,7 @@ static void doMergeIntervalAggImpl(SOperatorInfo* pOperatorInfo, SResultRowInfo* doWindowBorderInterpolation(iaInfo, pBlock, pResult, &nextWin, startPos, forwardRows, pExprSup); updateTimeWindowInfo(&iaInfo->twAggSup.timeWindowData, &nextWin, true); - doApplyFunctions(pTaskInfo, pExprSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, forwardRows, + applyAggFunctionOnPartialTuples(pTaskInfo, pExprSup->pCtx, &iaInfo->twAggSup.timeWindowData, startPos, forwardRows, pBlock->info.rows, numOfOutput); doCloseWindow(pResultRowInfo, iaInfo, pResult); @@ -5178,7 +4656,7 @@ SOperatorInfo* createMergeIntervalOperatorInfo(SOperatorInfo* downstream, SMerge goto _error; } - SSDataBlock* pResBlock = createResDataBlock(pIntervalPhyNode->window.node.pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pIntervalPhyNode->window.node.pOutputDataBlockDesc); initBasicInfo(&pIntervalInfo->binfo, pResBlock); initExecTimeWindowInfo(&pIntervalInfo->twAggSup.timeWindowData, &pIntervalInfo->win); @@ -5334,7 +4812,7 @@ SOperatorInfo* createStreamIntervalOperatorInfo(SOperatorInfo* downstream, SPhys SExprInfo* pExprInfo = createExprInfo(pIntervalPhyNode->window.pFuncs, NULL, &numOfCols); ASSERT(numOfCols > 0); - SSDataBlock* pResBlock = createResDataBlock(pPhyNode->pOutputDataBlockDesc); + SSDataBlock* pResBlock = createDataBlockFromDescNode(pPhyNode->pOutputDataBlockDesc); SInterval interval = { .interval = pIntervalPhyNode->interval, .sliding = pIntervalPhyNode->sliding, diff --git a/source/libs/function/src/builtinsimpl.c b/source/libs/function/src/builtinsimpl.c index 640baf4f94..e572ce7a40 100644 --- a/source/libs/function/src/builtinsimpl.c +++ b/source/libs/function/src/builtinsimpl.c @@ -729,375 +729,6 @@ bool getSumFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { return true; } -//int32_t getAvgInfoSize() { return (int32_t)sizeof(SAvgRes); } -// -//bool getAvgFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { -// pEnv->calcMemSize = sizeof(SAvgRes); -// return true; -//} -// -//bool avgFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResultInfo) { -// if (!functionSetup(pCtx, pResultInfo)) { -// return false; -// } -// -// SAvgRes* pRes = GET_ROWCELL_INTERBUF(pResultInfo); -// memset(pRes, 0, sizeof(SAvgRes)); -// return true; -//} - -//int32_t avgFunction(SqlFunctionCtx* pCtx) { -// int32_t numOfElem = 0; -// -// SInputColumnInfoData* pInput = &pCtx->input; -// SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; -// int32_t type = pInput->pData[0]->info.type; -// -// SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); -// pAvgRes->type = type; -// -// // computing based on the true data block -// SColumnInfoData* pCol = pInput->pData[0]; -// -// int32_t start = pInput->startRowIndex; -// int32_t numOfRows = pInput->numOfRows; -// -// if (IS_NULL_TYPE(type)) { -// numOfElem = 0; -// goto _avg_over; -// } -// -// if (pInput->colDataSMAIsSet) { -// numOfElem = numOfRows - pAgg->numOfNull; -// ASSERT(numOfElem >= 0); -// -// pAvgRes->count += numOfElem; -// if (IS_SIGNED_NUMERIC_TYPE(type)) { -// pAvgRes->sum.isum += pAgg->sum; -// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { -// pAvgRes->sum.usum += pAgg->sum; -// } else if (IS_FLOAT_TYPE(type)) { -// pAvgRes->sum.dsum += GET_DOUBLE_VAL((const char*)&(pAgg->sum)); -// } -// } else { // computing based on the true data block -// switch (type) { -// case TSDB_DATA_TYPE_TINYINT: { -// int8_t* plist = (int8_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.isum += plist[i]; -// } -// -// break; -// } -// -// case TSDB_DATA_TYPE_SMALLINT: { -// int16_t* plist = (int16_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.isum += plist[i]; -// } -// break; -// } -// -// case TSDB_DATA_TYPE_INT: { -// int32_t* plist = (int32_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.isum += plist[i]; -// } -// -// break; -// } -// -// case TSDB_DATA_TYPE_BIGINT: { -// int64_t* plist = (int64_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.isum += plist[i]; -// } -// break; -// } -// -// case TSDB_DATA_TYPE_UTINYINT: { -// uint8_t* plist = (uint8_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.usum += plist[i]; -// } -// -// break; -// } -// -// case TSDB_DATA_TYPE_USMALLINT: { -// uint16_t* plist = (uint16_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.usum += plist[i]; -// } -// break; -// } -// -// case TSDB_DATA_TYPE_UINT: { -// uint32_t* plist = (uint32_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.usum += plist[i]; -// } -// -// break; -// } -// -// case TSDB_DATA_TYPE_UBIGINT: { -// uint64_t* plist = (uint64_t*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.usum += plist[i]; -// } -// break; -// } -// -// case TSDB_DATA_TYPE_FLOAT: { -// float* plist = (float*)pCol->pData; -//// float val = 0; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.dsum += plist[i]; -// } -//// pAvgRes->sum.dsum = val; -// break; -// } -// -// case TSDB_DATA_TYPE_DOUBLE: { -// double* plist = (double*)pCol->pData; -// for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) { -// if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// numOfElem += 1; -// pAvgRes->count += 1; -// pAvgRes->sum.dsum += plist[i]; -// } -// break; -// } -// -// default: -// break; -// } -// } -// -//_avg_over: -// // data in the check operation are all null, not output -// SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); -// return TSDB_CODE_SUCCESS; -//} - -//static void avgTransferInfo(SAvgRes* pInput, SAvgRes* pOutput) { -// pOutput->type = pInput->type; -// if (IS_SIGNED_NUMERIC_TYPE(pOutput->type)) { -// pOutput->sum.isum += pInput->sum.isum; -// } else if (IS_UNSIGNED_NUMERIC_TYPE(pOutput->type)) { -// pOutput->sum.usum += pInput->sum.usum; -// } else { -// pOutput->sum.dsum += pInput->sum.dsum; -// } -// -// pOutput->count += pInput->count; -// -// return; -//} -// -//int32_t avgFunctionMerge(SqlFunctionCtx* pCtx) { -// SInputColumnInfoData* pInput = &pCtx->input; -// SColumnInfoData* pCol = pInput->pData[0]; -// ASSERT(pCol->info.type == TSDB_DATA_TYPE_BINARY); -// -// SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); -// -// int32_t start = pInput->startRowIndex; -// -// for (int32_t i = start; i < start + pInput->numOfRows; ++i) { -// char* data = colDataGetData(pCol, i); -// SAvgRes* pInputInfo = (SAvgRes*)varDataVal(data); -// avgTransferInfo(pInputInfo, pInfo); -// } -// -// SET_VAL(GET_RES_INFO(pCtx), 1, 1); -// -// return TSDB_CODE_SUCCESS; -//} -// -//int32_t avgInvertFunction(SqlFunctionCtx* pCtx) { -// int32_t numOfElem = 0; -// -// // Only the pre-computing information loaded and actual data does not loaded -// SInputColumnInfoData* pInput = &pCtx->input; -// int32_t type = pInput->pData[0]->info.type; -// -// SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); -// -// // computing based on the true data block -// SColumnInfoData* pCol = pInput->pData[0]; -// -// int32_t start = pInput->startRowIndex; -// int32_t numOfRows = pInput->numOfRows; -// -// switch (type) { -// case TSDB_DATA_TYPE_TINYINT: { -// LIST_AVG_N(pAvgRes->sum.isum, int8_t); -// break; -// } -// case TSDB_DATA_TYPE_SMALLINT: { -// LIST_AVG_N(pAvgRes->sum.isum, int16_t); -// break; -// } -// case TSDB_DATA_TYPE_INT: { -// LIST_AVG_N(pAvgRes->sum.isum, int32_t); -// break; -// } -// case TSDB_DATA_TYPE_BIGINT: { -// LIST_AVG_N(pAvgRes->sum.isum, int64_t); -// break; -// } -// case TSDB_DATA_TYPE_UTINYINT: { -// LIST_AVG_N(pAvgRes->sum.usum, uint8_t); -// break; -// } -// case TSDB_DATA_TYPE_USMALLINT: { -// LIST_AVG_N(pAvgRes->sum.usum, uint16_t); -// break; -// } -// case TSDB_DATA_TYPE_UINT: { -// LIST_AVG_N(pAvgRes->sum.usum, uint32_t); -// break; -// } -// case TSDB_DATA_TYPE_UBIGINT: { -// LIST_AVG_N(pAvgRes->sum.usum, uint64_t); -// break; -// } -// case TSDB_DATA_TYPE_FLOAT: { -// LIST_AVG_N(pAvgRes->sum.dsum, float); -// break; -// } -// case TSDB_DATA_TYPE_DOUBLE: { -// LIST_AVG_N(pAvgRes->sum.dsum, double); -// break; -// } -// default: -// break; -// } -// -// // data in the check operation are all null, not output -// SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1); -// return TSDB_CODE_SUCCESS; -//} -// -//int32_t avgCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx) { -// SResultRowEntryInfo* pDResInfo = GET_RES_INFO(pDestCtx); -// SAvgRes* pDBuf = GET_ROWCELL_INTERBUF(pDResInfo); -// -// SResultRowEntryInfo* pSResInfo = GET_RES_INFO(pSourceCtx); -// SAvgRes* pSBuf = GET_ROWCELL_INTERBUF(pSResInfo); -// int16_t type = pDBuf->type == TSDB_DATA_TYPE_NULL ? pSBuf->type : pDBuf->type; -// -// if (IS_SIGNED_NUMERIC_TYPE(type)) { -// pDBuf->sum.isum += pSBuf->sum.isum; -// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { -// pDBuf->sum.usum += pSBuf->sum.usum; -// } else { -// pDBuf->sum.dsum += pSBuf->sum.dsum; -// } -// pDBuf->count += pSBuf->count; -// -// return TSDB_CODE_SUCCESS; -//} -// -//int32_t avgFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { -// SInputColumnInfoData* pInput = &pCtx->input; -// -// SAvgRes* pAvgRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); -// int32_t type = pAvgRes->type; -// -// if (IS_SIGNED_NUMERIC_TYPE(type)) { -// pAvgRes->result = pAvgRes->sum.isum / ((double)pAvgRes->count); -// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { -// pAvgRes->result = pAvgRes->sum.usum / ((double)pAvgRes->count); -// } else { -// pAvgRes->result = pAvgRes->sum.dsum / ((double)pAvgRes->count); -// } -// -// // check for overflow -// if (isinf(pAvgRes->result) || isnan(pAvgRes->result)) { -// GET_RES_INFO(pCtx)->numOfRes = 0; -// } -// -// return functionFinalize(pCtx, pBlock); -//} -// -//int32_t avgPartialFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) { -// SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); -// SAvgRes* pInfo = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx)); -// int32_t resultBytes = getAvgInfoSize(); -// char* res = taosMemoryCalloc(resultBytes + VARSTR_HEADER_SIZE, sizeof(char)); -// -// memcpy(varDataVal(res), pInfo, resultBytes); -// varDataSetLen(res, resultBytes); -// -// int32_t slotId = pCtx->pExpr->base.resSchema.slotId; -// SColumnInfoData* pCol = taosArrayGet(pBlock->pDataBlock, slotId); -// -// colDataAppend(pCol, pBlock->info.rows, res, false); -// -// taosMemoryFree(res); -// return pResInfo->numOfRes; -//} - EFuncDataRequired statisDataRequired(SFunctionNode* pFunc, STimeWindow* pTimeWindow) { return FUNC_DATA_REQUIRED_SMA_LOAD; } @@ -1121,530 +752,6 @@ bool getMinmaxFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) { return true; } -//static STuplePos saveTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, -// const STupleKey* pKey); -//static int32_t updateTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos); -//static const char* loadTupleData(SqlFunctionCtx* pCtx, const STuplePos* pPos); - -//static int32_t findRowIndex(int32_t start, int32_t num, SColumnInfoData* pCol, const char* tval) { -// // the data is loaded, not only the block SMA value -// for (int32_t i = start; i < num + start; ++i) { -// char* p = colDataGetData(pCol, i); -// if (memcmp((void*)tval, p, pCol->info.bytes) == 0) { -// return i; -// } -// } -// -// // if reach here means real data of block SMA is not set in pCtx->input. -// return -1; -//} - -//int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) { -// int32_t numOfElems = 0; -// -// SInputColumnInfoData* pInput = &pCtx->input; -// SColumnDataAgg* pAgg = pInput->pColumnDataAgg[0]; -// -// SColumnInfoData* pCol = pInput->pData[0]; -// int32_t type = pCol->info.type; -// -// SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx); -// SMinmaxResInfo* pBuf = GET_ROWCELL_INTERBUF(pResInfo); -// pBuf->type = type; -// -// if (IS_NULL_TYPE(type)) { -// numOfElems = 0; -// goto _min_max_over; -// } -// -// // data in current data block are qualified to the query -// if (pInput->colDataSMAIsSet) { -// numOfElems = pInput->numOfRows - pAgg->numOfNull; -// ASSERT(pInput->numOfRows == pInput->totalRows && numOfElems >= 0); -// if (numOfElems == 0) { -// return numOfElems; -// } -// -// void* tval = NULL; -// int16_t index = 0; -// -// if (isMinFunc) { -// tval = &pInput->pColumnDataAgg[0]->min; -// } else { -// tval = &pInput->pColumnDataAgg[0]->max; -// } -// -// if (!pBuf->assign) { -// pBuf->v = *(int64_t*)tval; -// if (pCtx->subsidiaries.num > 0) { -// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); -// if (index >= 0) { -// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); -// } -// } -// } else { -// if (IS_SIGNED_NUMERIC_TYPE(type)) { -// int64_t prev = 0; -// GET_TYPED_DATA(prev, int64_t, type, &pBuf->v); -// -// int64_t val = GET_INT64_VAL(tval); -// if ((prev < val) ^ isMinFunc) { -// *(int64_t*)&pBuf->v = val; -// if (pCtx->subsidiaries.num > 0) { -// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); -// if (index >= 0) { -// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); -// } -// } -// } -// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { -// uint64_t prev = 0; -// GET_TYPED_DATA(prev, uint64_t, type, &pBuf->v); -// -// uint64_t val = GET_UINT64_VAL(tval); -// if ((prev < val) ^ isMinFunc) { -// *(uint64_t*)&pBuf->v = val; -// if (pCtx->subsidiaries.num > 0) { -// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); -// if (index >= 0) { -// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); -// } -// } -// } -// } else if (type == TSDB_DATA_TYPE_DOUBLE) { -// double prev = 0; -// GET_TYPED_DATA(prev, double, type, &pBuf->v); -// -// double val = GET_DOUBLE_VAL(tval); -// if ((prev < val) ^ isMinFunc) { -// *(double*)&pBuf->v = val; -// if (pCtx->subsidiaries.num > 0) { -// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); -// if (index >= 0) { -// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); -// } -// } -// } -// } else if (type == TSDB_DATA_TYPE_FLOAT) { -// float prev = 0; -// GET_TYPED_DATA(prev, float, type, &pBuf->v); -// -// float val = GET_DOUBLE_VAL(tval); -// if ((prev < val) ^ isMinFunc) { -// *(float*)&pBuf->v = val; -// } -// -// if (pCtx->subsidiaries.num > 0) { -// index = findRowIndex(pInput->startRowIndex, pInput->numOfRows, pCol, tval); -// if (index >= 0) { -// pBuf->tuplePos = saveTupleData(pCtx, index, pCtx->pSrcBlock, NULL); -// } -// } -// } -// } -// -// pBuf->assign = true; -// return numOfElems; -// } -// -// int32_t start = pInput->startRowIndex; -// int32_t numOfRows = pInput->numOfRows; -// -// if (IS_SIGNED_NUMERIC_TYPE(type) || type == TSDB_DATA_TYPE_BOOL) { -// if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) { -// int8_t* pData = (int8_t*)pCol->pData; -// int8_t* val = (int8_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } else if (type == TSDB_DATA_TYPE_SMALLINT) { -// int16_t* pData = (int16_t*)pCol->pData; -// int16_t* val = (int16_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } else if (type == TSDB_DATA_TYPE_INT) { -// int32_t* pData = (int32_t*)pCol->pData; -// int32_t* val = (int32_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } else if (type == TSDB_DATA_TYPE_BIGINT) { -// int64_t* pData = (int64_t*)pCol->pData; -// int64_t* val = (int64_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } -// } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) { -// if (type == TSDB_DATA_TYPE_UTINYINT) { -// uint8_t* pData = (uint8_t*)pCol->pData; -// uint8_t* val = (uint8_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } else if (type == TSDB_DATA_TYPE_USMALLINT) { -// uint16_t* pData = (uint16_t*)pCol->pData; -// uint16_t* val = (uint16_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } else if (type == TSDB_DATA_TYPE_UINT) { -// uint32_t* pData = (uint32_t*)pCol->pData; -// uint32_t* val = (uint32_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } else if (type == TSDB_DATA_TYPE_UBIGINT) { -// uint64_t* pData = (uint64_t*)pCol->pData; -// uint64_t* val = (uint64_t*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } -// } else if (type == TSDB_DATA_TYPE_DOUBLE) { -// double* pData = (double*)pCol->pData; -// double* val = (double*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -// // ignore the equivalent data value -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } else if (type == TSDB_DATA_TYPE_FLOAT) { -// float* pData = (float*)pCol->pData; -// float* val = (float*)&pBuf->v; -// -// for (int32_t i = start; i < start + numOfRows; ++i) { -// if ((pCol->hasNull) && colDataIsNull_f(pCol->nullbitmap, i)) { -// continue; -// } -// -// if (!pBuf->assign) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// pBuf->tuplePos = saveTupleData(pCtx, i, pCtx->pSrcBlock, NULL); -// } -// pBuf->assign = true; -// } else { -//#if 0 -// if ((*val) == pData[i]) { -// continue; -// } -// -// if ((*val < pData[i]) ^ isMinFunc) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -//#endif -// // NOTE: An faster version to avoid one additional comparison with FPU. -// if (isMinFunc) { // min -// if (*val > pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } else { // max -// if (*val < pData[i]) { -// *val = pData[i]; -// if (pCtx->subsidiaries.num > 0) { -// updateTupleData(pCtx, i, pCtx->pSrcBlock, &pBuf->tuplePos); -// } -// } -// } -// } -// -// numOfElems += 1; -// } -// } -// -//_min_max_over: -// if (numOfElems == 0 && pCtx->subsidiaries.num > 0 && !pBuf->nullTupleSaved) { -// pBuf->nullTuplePos = saveTupleData(pCtx, pInput->startRowIndex, pCtx->pSrcBlock, NULL); -// pBuf->nullTupleSaved = true; -// } -// return numOfElems; -//} - int32_t minFunction(SqlFunctionCtx* pCtx) { int32_t numOfElems = doMinMaxHelper(pCtx, 1); SET_VAL(GET_RES_INFO(pCtx), numOfElems, 1); From 87d9836a756d154e78501076174cb206aaf57009 Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Mon, 28 Nov 2022 00:28:29 +0800 Subject: [PATCH 82/83] refactor: do some internal refactor. --- source/libs/executor/src/timesliceoperator.c | 587 +++++++++++++++++++ 1 file changed, 587 insertions(+) create mode 100644 source/libs/executor/src/timesliceoperator.c diff --git a/source/libs/executor/src/timesliceoperator.c b/source/libs/executor/src/timesliceoperator.c new file mode 100644 index 0000000000..d8cef86971 --- /dev/null +++ b/source/libs/executor/src/timesliceoperator.c @@ -0,0 +1,587 @@ +/* + * Copyright (c) 2019 TAOS Data, Inc. + * + * This program is free software: you can use, redistribute, and/or modify + * it under the terms of the GNU Affero General Public License, version 3 + * or later ("AGPL"), as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. + * + * You should have received a copy of the GNU Affero General Public License + * along with this program. If not, see . + */ +#include "executorimpl.h" +#include "filter.h" +#include "function.h" +#include "functionMgt.h" +#include "tcommon.h" +#include "tcompare.h" +#include "tdatablock.h" +#include "tfill.h" +#include "ttime.h" + +typedef struct STimeSliceOperatorInfo { + SSDataBlock* pRes; + STimeWindow win; + SInterval interval; + int64_t current; + SArray* pPrevRow; // SArray + SArray* pNextRow; // SArray + SArray* pLinearInfo; // SArray + bool isPrevRowSet; + bool isNextRowSet; + int32_t fillType; // fill type + SColumn tsCol; // primary timestamp column + SExprSupp scalarSup; // scalar calculation + struct SFillColInfo* pFillColInfo; // fill column info +} STimeSliceOperatorInfo; + +static void destroyTimeSliceOperatorInfo(void* param); + +static void doKeepPrevRows(STimeSliceOperatorInfo* pSliceInfo, const SSDataBlock* pBlock, int32_t rowIndex) { + int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); + for (int32_t i = 0; i < numOfCols; ++i) { + SColumnInfoData* pColInfoData = taosArrayGet(pBlock->pDataBlock, i); + + SGroupKeys* pkey = taosArrayGet(pSliceInfo->pPrevRow, i); + if (!colDataIsNull_s(pColInfoData, rowIndex)) { + pkey->isNull = false; + char* val = colDataGetData(pColInfoData, rowIndex); + if (!IS_VAR_DATA_TYPE(pkey->type)) { + memcpy(pkey->pData, val, pkey->bytes); + } else { + memcpy(pkey->pData, val, varDataLen(val)); + } + } else { + pkey->isNull = true; + } + } + + pSliceInfo->isPrevRowSet = true; +} + +static void doKeepNextRows(STimeSliceOperatorInfo* pSliceInfo, const SSDataBlock* pBlock, int32_t rowIndex) { + int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); + for (int32_t i = 0; i < numOfCols; ++i) { + SColumnInfoData* pColInfoData = taosArrayGet(pBlock->pDataBlock, i); + + SGroupKeys* pkey = taosArrayGet(pSliceInfo->pNextRow, i); + if (!colDataIsNull_s(pColInfoData, rowIndex)) { + pkey->isNull = false; + char* val = colDataGetData(pColInfoData, rowIndex); + if (!IS_VAR_DATA_TYPE(pkey->type)) { + memcpy(pkey->pData, val, pkey->bytes); + } else { + memcpy(pkey->pData, val, varDataLen(val)); + } + } else { + pkey->isNull = true; + } + } + + pSliceInfo->isNextRowSet = true; +} + +static void doKeepLinearInfo(STimeSliceOperatorInfo* pSliceInfo, const SSDataBlock* pBlock, int32_t rowIndex) { + int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); + for (int32_t i = 0; i < numOfCols; ++i) { + SColumnInfoData* pColInfoData = taosArrayGet(pBlock->pDataBlock, i); + SColumnInfoData* pTsCol = taosArrayGet(pBlock->pDataBlock, pSliceInfo->tsCol.slotId); + SFillLinearInfo* pLinearInfo = taosArrayGet(pSliceInfo->pLinearInfo, i); + + // null value is represented by using key = INT64_MIN for now. + // TODO: optimize to ignore null values for linear interpolation. + if (!pLinearInfo->isStartSet) { + if (!colDataIsNull_s(pColInfoData, rowIndex)) { + pLinearInfo->start.key = *(int64_t*)colDataGetData(pTsCol, rowIndex); + memcpy(pLinearInfo->start.val, colDataGetData(pColInfoData, rowIndex), pLinearInfo->bytes); + } + pLinearInfo->isStartSet = true; + } else if (!pLinearInfo->isEndSet) { + if (!colDataIsNull_s(pColInfoData, rowIndex)) { + pLinearInfo->end.key = *(int64_t*)colDataGetData(pTsCol, rowIndex); + memcpy(pLinearInfo->end.val, colDataGetData(pColInfoData, rowIndex), pLinearInfo->bytes); + } + pLinearInfo->isEndSet = true; + } else { + pLinearInfo->start.key = pLinearInfo->end.key; + memcpy(pLinearInfo->start.val, pLinearInfo->end.val, pLinearInfo->bytes); + + if (!colDataIsNull_s(pColInfoData, rowIndex)) { + pLinearInfo->end.key = *(int64_t*)colDataGetData(pTsCol, rowIndex); + memcpy(pLinearInfo->end.val, colDataGetData(pColInfoData, rowIndex), pLinearInfo->bytes); + } else { + pLinearInfo->end.key = INT64_MIN; + } + } + } + +} + +static bool genInterpolationResult(STimeSliceOperatorInfo* pSliceInfo, SExprSupp* pExprSup, SSDataBlock* pResBlock, bool beforeTs) { + int32_t rows = pResBlock->info.rows; + blockDataEnsureCapacity(pResBlock, rows + 1); + // todo set the correct primary timestamp column + + // output the result + bool hasInterp = true; + for (int32_t j = 0; j < pExprSup->numOfExprs; ++j) { + SExprInfo* pExprInfo = &pExprSup->pExprInfo[j]; + + int32_t dstSlot = pExprInfo->base.resSchema.slotId; + SColumnInfoData* pDst = taosArrayGet(pResBlock->pDataBlock, dstSlot); + + if (IS_TIMESTAMP_TYPE(pExprInfo->base.resSchema.type)) { + colDataAppend(pDst, rows, (char*)&pSliceInfo->current, false); + continue; + } + + int32_t srcSlot = pExprInfo->base.pParam[0].pCol->slotId; + switch (pSliceInfo->fillType) { + case TSDB_FILL_NULL: { + colDataAppendNULL(pDst, rows); + break; + } + + case TSDB_FILL_SET_VALUE: { + SVariant* pVar = &pSliceInfo->pFillColInfo[j].fillVal; + + if (pDst->info.type == TSDB_DATA_TYPE_FLOAT) { + float v = 0; + GET_TYPED_DATA(v, float, pVar->nType, &pVar->i); + colDataAppend(pDst, rows, (char*)&v, false); + } else if (pDst->info.type == TSDB_DATA_TYPE_DOUBLE) { + double v = 0; + GET_TYPED_DATA(v, double, pVar->nType, &pVar->i); + colDataAppend(pDst, rows, (char*)&v, false); + } else if (IS_SIGNED_NUMERIC_TYPE(pDst->info.type)) { + int64_t v = 0; + GET_TYPED_DATA(v, int64_t, pVar->nType, &pVar->i); + colDataAppend(pDst, rows, (char*)&v, false); + } + break; + } + + case TSDB_FILL_LINEAR: { + SFillLinearInfo* pLinearInfo = taosArrayGet(pSliceInfo->pLinearInfo, srcSlot); + + SPoint start = pLinearInfo->start; + SPoint end = pLinearInfo->end; + SPoint current = {.key = pSliceInfo->current}; + + // do not interpolate before ts range, only increate pSliceInfo->current + if (beforeTs && !pLinearInfo->isEndSet) { + return true; + } + + if (!pLinearInfo->isStartSet || !pLinearInfo->isEndSet) { + hasInterp = false; + break; + } + + if (start.key == INT64_MIN || end.key == INT64_MIN) { + colDataAppendNULL(pDst, rows); + break; + } + + current.val = taosMemoryCalloc(pLinearInfo->bytes, 1); + taosGetLinearInterpolationVal(¤t, pLinearInfo->type, &start, &end, pLinearInfo->type); + colDataAppend(pDst, rows, (char*)current.val, false); + + taosMemoryFree(current.val); + break; + } + case TSDB_FILL_PREV: { + if (!pSliceInfo->isPrevRowSet) { + hasInterp = false; + break; + } + + SGroupKeys* pkey = taosArrayGet(pSliceInfo->pPrevRow, srcSlot); + if (pkey->isNull == false) { + colDataAppend(pDst, rows, pkey->pData, false); + } else { + colDataAppendNULL(pDst, rows); + } + break; + } + + case TSDB_FILL_NEXT: { + if (!pSliceInfo->isNextRowSet) { + hasInterp = false; + break; + } + + SGroupKeys* pkey = taosArrayGet(pSliceInfo->pNextRow, srcSlot); + if (pkey->isNull == false) { + colDataAppend(pDst, rows, pkey->pData, false); + } else { + colDataAppendNULL(pDst, rows); + } + break; + } + + case TSDB_FILL_NONE: + default: + break; + } + } + + if (hasInterp) { + pResBlock->info.rows += 1; + } + + return hasInterp; +} + +static void addCurrentRowToResult(STimeSliceOperatorInfo* pSliceInfo, SExprSupp* pExprSup, SSDataBlock* pResBlock, + SSDataBlock* pSrcBlock, int32_t index) { + blockDataEnsureCapacity(pResBlock, pResBlock->info.rows + 1); + for (int32_t j = 0; j < pExprSup->numOfExprs; ++j) { + SExprInfo* pExprInfo = &pExprSup->pExprInfo[j]; + + int32_t dstSlot = pExprInfo->base.resSchema.slotId; + SColumnInfoData* pDst = taosArrayGet(pResBlock->pDataBlock, dstSlot); + + if (IS_TIMESTAMP_TYPE(pExprInfo->base.resSchema.type)) { + colDataAppend(pDst, pResBlock->info.rows, (char*)&pSliceInfo->current, false); + } else { + int32_t srcSlot = pExprInfo->base.pParam[0].pCol->slotId; + SColumnInfoData* pSrc = taosArrayGet(pSrcBlock->pDataBlock, srcSlot); + + if (colDataIsNull_s(pSrc, index)) { + colDataAppendNULL(pDst, pResBlock->info.rows); + continue; + } + + char* v = colDataGetData(pSrc, index); + colDataAppend(pDst, pResBlock->info.rows, v, false); + } + } + + pResBlock->info.rows += 1; + return; +} + + +static int32_t initPrevRowsKeeper(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { + if (pInfo->pPrevRow != NULL) { + return TSDB_CODE_SUCCESS; + } + + pInfo->pPrevRow = taosArrayInit(4, sizeof(SGroupKeys)); + if (pInfo->pPrevRow == NULL) { + return TSDB_CODE_OUT_OF_MEMORY; + } + + int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); + for (int32_t i = 0; i < numOfCols; ++i) { + SColumnInfoData* pColInfo = taosArrayGet(pBlock->pDataBlock, i); + + SGroupKeys key = {0}; + key.bytes = pColInfo->info.bytes; + key.type = pColInfo->info.type; + key.isNull = false; + key.pData = taosMemoryCalloc(1, pColInfo->info.bytes); + taosArrayPush(pInfo->pPrevRow, &key); + } + + pInfo->isPrevRowSet = false; + + return TSDB_CODE_SUCCESS; +} + +static int32_t initNextRowsKeeper(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { + if (pInfo->pNextRow != NULL) { + return TSDB_CODE_SUCCESS; + } + + pInfo->pNextRow = taosArrayInit(4, sizeof(SGroupKeys)); + if (pInfo->pNextRow == NULL) { + return TSDB_CODE_OUT_OF_MEMORY; + } + + int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); + for (int32_t i = 0; i < numOfCols; ++i) { + SColumnInfoData* pColInfo = taosArrayGet(pBlock->pDataBlock, i); + + SGroupKeys key = {0}; + key.bytes = pColInfo->info.bytes; + key.type = pColInfo->info.type; + key.isNull = false; + key.pData = taosMemoryCalloc(1, pColInfo->info.bytes); + taosArrayPush(pInfo->pNextRow, &key); + } + + pInfo->isNextRowSet = false; + + return TSDB_CODE_SUCCESS; +} + +static int32_t initFillLinearInfo(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { + if (pInfo->pLinearInfo != NULL) { + return TSDB_CODE_SUCCESS; + } + + pInfo->pLinearInfo = taosArrayInit(4, sizeof(SFillLinearInfo)); + if (pInfo->pLinearInfo == NULL) { + return TSDB_CODE_OUT_OF_MEMORY; + } + + int32_t numOfCols = taosArrayGetSize(pBlock->pDataBlock); + for (int32_t i = 0; i < numOfCols; ++i) { + SColumnInfoData* pColInfo = taosArrayGet(pBlock->pDataBlock, i); + + SFillLinearInfo linearInfo = {0}; + linearInfo.start.key = INT64_MIN; + linearInfo.end.key = INT64_MIN; + linearInfo.start.val = taosMemoryCalloc(1, pColInfo->info.bytes); + linearInfo.end.val = taosMemoryCalloc(1, pColInfo->info.bytes); + linearInfo.isStartSet = false; + linearInfo.isEndSet = false; + linearInfo.type = pColInfo->info.type; + linearInfo.bytes = pColInfo->info.bytes; + taosArrayPush(pInfo->pLinearInfo, &linearInfo); + } + + return TSDB_CODE_SUCCESS; +} + +static int32_t initKeeperInfo(STimeSliceOperatorInfo* pInfo, SSDataBlock* pBlock) { + int32_t code; + code = initPrevRowsKeeper(pInfo, pBlock); + if (code != TSDB_CODE_SUCCESS) { + return TSDB_CODE_FAILED; + } + + code = initNextRowsKeeper(pInfo, pBlock); + if (code != TSDB_CODE_SUCCESS) { + return TSDB_CODE_FAILED; + } + + code = initFillLinearInfo(pInfo, pBlock); + if (code != TSDB_CODE_SUCCESS) { + return TSDB_CODE_FAILED; + } + + return TSDB_CODE_SUCCESS; +} + +static SSDataBlock* doTimeslice(SOperatorInfo* pOperator) { + if (pOperator->status == OP_EXEC_DONE) { + return NULL; + } + + SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo; + + STimeSliceOperatorInfo* pSliceInfo = pOperator->info; + SSDataBlock* pResBlock = pSliceInfo->pRes; + SExprSupp* pSup = &pOperator->exprSupp; + + int32_t order = TSDB_ORDER_ASC; + SInterval* pInterval = &pSliceInfo->interval; + SOperatorInfo* downstream = pOperator->pDownstream[0]; + + blockDataCleanup(pResBlock); + + while (1) { + SSDataBlock* pBlock = downstream->fpSet.getNextFn(downstream); + if (pBlock == NULL) { + break; + } + + int32_t code = initKeeperInfo(pSliceInfo, pBlock); + if (code != TSDB_CODE_SUCCESS) { + T_LONG_JMP(pTaskInfo->env, code); + } + + // the pDataBlock are always the same one, no need to call this again + setInputDataBlock(pSup, pBlock, order, MAIN_SCAN, true); + + SColumnInfoData* pTsCol = taosArrayGet(pBlock->pDataBlock, pSliceInfo->tsCol.slotId); + for (int32_t i = 0; i < pBlock->info.rows; ++i) { + int64_t ts = *(int64_t*)colDataGetData(pTsCol, i); + + if (pSliceInfo->current > pSliceInfo->win.ekey) { + setOperatorCompleted(pOperator); + break; + } + + if (ts == pSliceInfo->current) { + addCurrentRowToResult(pSliceInfo, &pOperator->exprSupp, pResBlock, pBlock, i); + + doKeepPrevRows(pSliceInfo, pBlock, i); + doKeepLinearInfo(pSliceInfo, pBlock, i); + + pSliceInfo->current = + taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); + if (pSliceInfo->current > pSliceInfo->win.ekey) { + setOperatorCompleted(pOperator); + break; + } + } else if (ts < pSliceInfo->current) { + // in case of interpolation window starts and ends between two datapoints, fill(prev) need to interpolate + doKeepPrevRows(pSliceInfo, pBlock, i); + doKeepLinearInfo(pSliceInfo, pBlock, i); + + if (i < pBlock->info.rows - 1) { + // in case of interpolation window starts and ends between two datapoints, fill(next) need to interpolate + doKeepNextRows(pSliceInfo, pBlock, i + 1); + int64_t nextTs = *(int64_t*)colDataGetData(pTsCol, i + 1); + if (nextTs > pSliceInfo->current) { + while (pSliceInfo->current < nextTs && pSliceInfo->current <= pSliceInfo->win.ekey) { + if (!genInterpolationResult(pSliceInfo, &pOperator->exprSupp, pResBlock, false) && pSliceInfo->fillType == TSDB_FILL_LINEAR) { + break; + } else { + pSliceInfo->current = + taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); + } + } + + if (pSliceInfo->current > pSliceInfo->win.ekey) { + setOperatorCompleted(pOperator); + break; + } + } else { + // ignore current row, and do nothing + } + } else { // it is the last row of current block + doKeepPrevRows(pSliceInfo, pBlock, i); + } + } else { // ts > pSliceInfo->current + // in case of interpolation window starts and ends between two datapoints, fill(next) need to interpolate + doKeepNextRows(pSliceInfo, pBlock, i); + doKeepLinearInfo(pSliceInfo, pBlock, i); + + while (pSliceInfo->current < ts && pSliceInfo->current <= pSliceInfo->win.ekey) { + if (!genInterpolationResult(pSliceInfo, &pOperator->exprSupp, pResBlock, true) && pSliceInfo->fillType == TSDB_FILL_LINEAR) { + break; + } else { + pSliceInfo->current = + taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); + } + } + + // add current row if timestamp match + if (ts == pSliceInfo->current && pSliceInfo->current <= pSliceInfo->win.ekey) { + addCurrentRowToResult(pSliceInfo, &pOperator->exprSupp, pResBlock, pBlock, i); + doKeepPrevRows(pSliceInfo, pBlock, i); + + pSliceInfo->current = + taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); + } + + if (pSliceInfo->current > pSliceInfo->win.ekey) { + setOperatorCompleted(pOperator); + break; + } + } + } + } + + // check if need to interpolate after last datablock + // except for fill(next), fill(linear) + while (pSliceInfo->current <= pSliceInfo->win.ekey && pSliceInfo->fillType != TSDB_FILL_NEXT && + pSliceInfo->fillType != TSDB_FILL_LINEAR) { + genInterpolationResult(pSliceInfo, &pOperator->exprSupp, pResBlock, false); + pSliceInfo->current = + taosTimeAdd(pSliceInfo->current, pInterval->interval, pInterval->intervalUnit, pInterval->precision); + } + + // restore the value + setTaskStatus(pOperator->pTaskInfo, TASK_COMPLETED); + if (pResBlock->info.rows == 0) { + pOperator->status = OP_EXEC_DONE; + } + + return pResBlock->info.rows == 0 ? NULL : pResBlock; +} + +SOperatorInfo* createTimeSliceOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo) { + STimeSliceOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(STimeSliceOperatorInfo)); + SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo)); + if (pOperator == NULL || pInfo == NULL) { + goto _error; + } + + SInterpFuncPhysiNode* pInterpPhyNode = (SInterpFuncPhysiNode*)pPhyNode; + SExprSupp* pSup = &pOperator->exprSupp; + + int32_t numOfExprs = 0; + SExprInfo* pExprInfo = createExprInfo(pInterpPhyNode->pFuncs, NULL, &numOfExprs); + int32_t code = initExprSupp(pSup, pExprInfo, numOfExprs); + if (code != TSDB_CODE_SUCCESS) { + goto _error; + } + + if (pInterpPhyNode->pExprs != NULL) { + int32_t num = 0; + SExprInfo* pScalarExprInfo = createExprInfo(pInterpPhyNode->pExprs, NULL, &num); + code = initExprSupp(&pInfo->scalarSup, pScalarExprInfo, num); + if (code != TSDB_CODE_SUCCESS) { + goto _error; + } + } + + pInfo->tsCol = extractColumnFromColumnNode((SColumnNode*)pInterpPhyNode->pTimeSeries); + pInfo->fillType = convertFillType(pInterpPhyNode->fillMode); + initResultSizeInfo(&pOperator->resultInfo, 4096); + + pInfo->pFillColInfo = createFillColInfo(pExprInfo, numOfExprs, NULL, 0, (SNodeListNode*)pInterpPhyNode->pFillValues); + pInfo->pLinearInfo = NULL; + pInfo->pRes = createDataBlockFromDescNode(pPhyNode->pOutputDataBlockDesc); + pInfo->win = pInterpPhyNode->timeRange; + pInfo->interval.interval = pInterpPhyNode->interval; + pInfo->current = pInfo->win.skey; + + if (downstream->operatorType == QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN) { + STableScanInfo* pScanInfo = (STableScanInfo*)downstream->info; + pScanInfo->base.cond.twindows = pInfo->win; + pScanInfo->base.cond.type = TIMEWINDOW_RANGE_EXTERNAL; + } + + setOperatorInfo(pOperator, "TimeSliceOperator", QUERY_NODE_PHYSICAL_PLAN_INTERP_FUNC, false, OP_NOT_OPENED, pInfo, + pTaskInfo); + pOperator->fpSet = createOperatorFpSet(operatorDummyOpenFn, doTimeslice, NULL, destroyTimeSliceOperatorInfo, NULL); + + blockDataEnsureCapacity(pInfo->pRes, pOperator->resultInfo.capacity); + + code = appendDownstream(pOperator, &downstream, 1); + return pOperator; + + _error: + taosMemoryFree(pInfo); + taosMemoryFree(pOperator); + pTaskInfo->code = TSDB_CODE_OUT_OF_MEMORY; + return NULL; +} + +void destroyTimeSliceOperatorInfo(void* param) { + STimeSliceOperatorInfo* pInfo = (STimeSliceOperatorInfo*)param; + + pInfo->pRes = blockDataDestroy(pInfo->pRes); + + for (int32_t i = 0; i < taosArrayGetSize(pInfo->pPrevRow); ++i) { + SGroupKeys* pKey = taosArrayGet(pInfo->pPrevRow, i); + taosMemoryFree(pKey->pData); + } + taosArrayDestroy(pInfo->pPrevRow); + + for (int32_t i = 0; i < taosArrayGetSize(pInfo->pNextRow); ++i) { + SGroupKeys* pKey = taosArrayGet(pInfo->pNextRow, i); + taosMemoryFree(pKey->pData); + } + taosArrayDestroy(pInfo->pNextRow); + + for (int32_t i = 0; i < taosArrayGetSize(pInfo->pLinearInfo); ++i) { + SFillLinearInfo* pKey = taosArrayGet(pInfo->pLinearInfo, i); + taosMemoryFree(pKey->start.val); + taosMemoryFree(pKey->end.val); + } + taosArrayDestroy(pInfo->pLinearInfo); + + taosMemoryFree(pInfo->pFillColInfo); + taosMemoryFreeClear(param); +} From d8da10f2bb697dddb79bcdf9958521d1e8e5114e Mon Sep 17 00:00:00 2001 From: Haojun Liao Date: Mon, 28 Nov 2022 00:51:18 +0800 Subject: [PATCH 83/83] refactor: do some internal refactor. --- source/libs/executor/inc/executil.h | 30 ++--- source/libs/executor/inc/executorimpl.h | 61 ++++------- source/libs/executor/src/executor.c | 17 +-- source/libs/executor/src/executorimpl.c | 103 +----------------- source/libs/executor/src/groupoperator.c | 2 +- source/libs/executor/src/projectoperator.c | 4 +- source/libs/executor/src/timewindowoperator.c | 35 ++---- 7 files changed, 56 insertions(+), 196 deletions(-) diff --git a/source/libs/executor/inc/executil.h b/source/libs/executor/inc/executil.h index d5366f1b7a..875528576d 100644 --- a/source/libs/executor/inc/executil.h +++ b/source/libs/executor/inc/executil.h @@ -38,16 +38,7 @@ memcpy((_k) + sizeof(uint64_t), (_ori), (_len)); \ } while (0) -#define SET_RES_EXT_WINDOW_KEY(_k, _ori, _len, _uid, _buf) \ - do { \ - assert(sizeof(_uid) == sizeof(uint64_t)); \ - *(void**)(_k) = (_buf); \ - *(uint64_t*)((_k) + POINTER_BYTES) = (_uid); \ - memcpy((_k) + POINTER_BYTES + sizeof(uint64_t), (_ori), (_len)); \ - } while (0) - #define GET_RES_WINDOW_KEY_LEN(_l) ((_l) + sizeof(uint64_t)) -#define GET_RES_EXT_WINDOW_KEY_LEN(_l) ((_l) + sizeof(uint64_t) + POINTER_BYTES) #define GET_TASKID(_t) (((SExecTaskInfo*)(_t))->id.str) @@ -104,16 +95,17 @@ int32_t createScanTableListInfo(SScanPhysiNode* pScanNode, SNodeList* pGroupTags STableListInfo* pTableListInfo, SNode* pTagCond, SNode* pTagIndexCond, SExecTaskInfo* pTaskInfo); STableListInfo* tableListCreate(); -void* tableListDestroy(STableListInfo* pTableListInfo); -void tableListClear(STableListInfo* pTableListInfo); -int32_t tableListGetOutputGroups(const STableListInfo* pTableList); -bool oneTableForEachGroup(const STableListInfo* pTableList); -uint64_t getTableGroupId(const STableListInfo* pTableList, uint64_t tableUid); -int32_t tableListAddTableInfo(STableListInfo* pTableList, uint64_t uid, uint64_t gid); -int32_t tableListGetGroupList(const STableListInfo* pTableList, int32_t ordinalIndex, STableKeyInfo** pKeyInfo, int32_t* num); -uint64_t tableListGetSize(const STableListInfo* pTableList); -uint64_t tableListGetSuid(const STableListInfo* pTableList); -STableKeyInfo* tableListGetInfo(const STableListInfo* pTableList, int32_t index); +void* tableListDestroy(STableListInfo* pTableListInfo); +void tableListClear(STableListInfo* pTableListInfo); +int32_t tableListGetOutputGroups(const STableListInfo* pTableList); +bool oneTableForEachGroup(const STableListInfo* pTableList); +uint64_t getTableGroupId(const STableListInfo* pTableList, uint64_t tableUid); +int32_t tableListAddTableInfo(STableListInfo* pTableList, uint64_t uid, uint64_t gid); +int32_t tableListGetGroupList(const STableListInfo* pTableList, int32_t ordinalIndex, STableKeyInfo** pKeyInfo, + int32_t* num); +uint64_t tableListGetSize(const STableListInfo* pTableList); +uint64_t tableListGetSuid(const STableListInfo* pTableList); +STableKeyInfo* tableListGetInfo(const STableListInfo* pTableList, int32_t index); size_t getResultRowSize(struct SqlFunctionCtx* pCtx, int32_t numOfOutput); void initResultRowInfo(SResultRowInfo* pResultRowInfo); diff --git a/source/libs/executor/inc/executorimpl.h b/source/libs/executor/inc/executorimpl.h index bd4472327c..8163217039 100644 --- a/source/libs/executor/inc/executorimpl.h +++ b/source/libs/executor/inc/executorimpl.h @@ -665,18 +665,25 @@ typedef struct SStreamFillOperatorInfo { SOperatorFpSet createOperatorFpSet(__optr_open_fn_t openFn, __optr_fn_t nextFn, __optr_fn_t cleanup, __optr_close_fn_t closeFn, __optr_explain_fn_t explain); - -int32_t operatorDummyOpenFn(SOperatorInfo* pOperator); -int32_t appendDownstream(SOperatorInfo* p, SOperatorInfo** pDownstream, int32_t num); +int32_t operatorDummyOpenFn(SOperatorInfo* pOperator); +int32_t appendDownstream(SOperatorInfo* p, SOperatorInfo** pDownstream, int32_t num); +void setOperatorCompleted(SOperatorInfo* pOperator); +void setOperatorInfo(SOperatorInfo* pOperator, const char* name, int32_t type, bool blocking, int32_t status, + void* pInfo, SExecTaskInfo* pTaskInfo); +void destroyOperatorInfo(SOperatorInfo* pOperator); void initBasicInfo(SOptrBasicInfo* pInfo, SSDataBlock* pBlock); void cleanupBasicInfo(SOptrBasicInfo* pInfo); + int32_t initExprSupp(SExprSupp* pSup, SExprInfo* pExprInfo, int32_t numOfExpr); void cleanupExprSupp(SExprSupp* pSup); + void destroyExprInfo(SExprInfo* pExpr, int32_t numOfExprs); -int32_t initAggInfo(SExprSupp* pSup, SAggSupporter* pAggSup, SExprInfo* pExprInfo, int32_t numOfCols, size_t keyBufSize, - const char* pkey); +int32_t initAggSup(SExprSupp* pSup, SAggSupporter* pAggSup, SExprInfo* pExprInfo, int32_t numOfCols, size_t keyBufSize, + const char* pkey); +void cleanupAggSup(SAggSupporter* pAggSup); + void initResultSizeInfo(SResultInfo* pResultInfo, int32_t numOfRows); void doBuildStreamResBlock(SOperatorInfo* pOperator, SOptrBasicInfo* pbInfo, SGroupResInfo* pGroupResInfo, @@ -702,14 +709,10 @@ int32_t getBufferPgSize(int32_t rowSize, uint32_t* defaultPgsz, uint32_t* defaul extern void doDestroyExchangeOperatorInfo(void* param); -void setOperatorCompleted(SOperatorInfo* pOperator); -void setOperatorInfo(SOperatorInfo* pOperator, const char* name, int32_t type, bool blocking, int32_t status, - void* pInfo, SExecTaskInfo* pTaskInfo); void doFilter(SSDataBlock* pBlock, SFilterInfo* pFilterInfo, SColMatchInfo* pColMatchInfo); int32_t addTagPseudoColumnData(SReadHandle* pHandle, const SExprInfo* pExpr, int32_t numOfExpr, SSDataBlock* pBlock, int32_t rows, const char* idStr, STableMetaCacheInfo* pCache); -void cleanupAggSup(SAggSupporter* pAggSup); void appendOneRowToDataBlock(SSDataBlock* pBlock, STupleHandle* pTupleHandle); void setTbNameColData(const SSDataBlock* pBlock, SColumnInfoData* pColInfoData, int32_t functionId, const char* name); @@ -724,6 +727,8 @@ SOperatorInfo* createExchangeOperatorInfo(void* pTransporter, SExchangePhysiNode SOperatorInfo* createTableScanOperatorInfo(STableScanPhysiNode* pTableScanNode, SReadHandle* pHandle, SExecTaskInfo* pTaskInfo); +SOperatorInfo* createTableMergeScanOperatorInfo(STableScanPhysiNode* pTableScanNode, SReadHandle* readHandle, SExecTaskInfo* pTaskInfo); + SOperatorInfo* createTagScanOperatorInfo(SReadHandle* pReadHandle, STagScanPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo); SOperatorInfo* createSysTableScanOperatorInfo(void* readHandle, SSystemTableScanPhysiNode* pScanPhyNode, const char* pUser, SExecTaskInfo* pTaskInfo); @@ -779,6 +784,8 @@ SOperatorInfo* createStreamIntervalOperatorInfo(SOperatorInfo* downstream, SPhys SOperatorInfo* createStreamStateAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo); SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFillPhysiNode* pPhyFillNode, SExecTaskInfo* pTaskInfo); + +SOperatorInfo* createGroupSortOperatorInfo(SOperatorInfo* downstream, SGroupSortPhysiNode* pSortPhyNode, SExecTaskInfo* pTaskInfo); // clang-format on int32_t projectApplyFunctions(SExprInfo* pExpr, SSDataBlock* pResult, SSDataBlock* pSrcBlock, SqlFunctionCtx* pCtx, @@ -786,38 +793,22 @@ int32_t projectApplyFunctions(SExprInfo* pExpr, SSDataBlock* pResult, SSDataBloc void setInputDataBlock(SExprSupp* pExprSupp, SSDataBlock* pBlock, int32_t order, int32_t scanFlag, bool createDummyCol); -bool isTaskKilled(SExecTaskInfo* pTaskInfo); int32_t checkForQueryBuf(size_t numOfTables); +bool isTaskKilled(SExecTaskInfo* pTaskInfo); void setTaskKilled(SExecTaskInfo* pTaskInfo); -void queryCostStatis(SExecTaskInfo* pTaskInfo); void doDestroyTask(SExecTaskInfo* pTaskInfo); -void destroyOperatorInfo(SOperatorInfo* pOperator); -int32_t getMaximumIdleDurationSec(); - -/* - * ops: root operator - * data: *data save the result of encode, need to be freed by caller - * length: *length save the length of *data - * nOptrWithVal: *nOptrWithVal save the number of optr with value - * return: result code, 0 means success - */ -int32_t encodeOperator(SOperatorInfo* ops, char** data, int32_t* length, int32_t* nOptrWithVal); - -/* - * ops: root operator, created by caller - * data: save the result of decode - * length: the length of data - * return: result code, 0 means success - */ -int32_t decodeOperator(SOperatorInfo* ops, const char* data, int32_t length); - void setTaskStatus(SExecTaskInfo* pTaskInfo, int8_t status); + int32_t createExecTaskInfoImpl(SSubplan* pPlan, SExecTaskInfo** pTaskInfo, SReadHandle* pHandle, uint64_t taskId, char* sql, EOPTR_EXEC_MODEL model); int32_t createDataSinkParam(SDataSinkNode* pNode, void** pParam, qTaskInfo_t* pTaskInfo, SReadHandle* readHandle); int32_t getOperatorExplainExecInfo(SOperatorInfo* operatorInfo, SArray* pExecInfoList); +void printTaskExecCostInLog(SExecTaskInfo* pTaskInfo); + +int32_t getMaximumIdleDurationSec(); + STimeWindow getActiveTimeWindow(SDiskbasedBuf* pBuf, SResultRowInfo* pResultRowInfo, int64_t ts, SInterval* pInterval, int32_t order); int32_t getNumOfRowsInTimeWindow(SDataBlockInfo* pDataBlockInfo, TSKEY* pPrimaryColumn, int32_t startPos, TSKEY ekey, @@ -840,15 +831,7 @@ void calBlockTbName(SStreamScanInfo* pInfo, SSDataBlock* pBlock); int32_t finalizeResultRows(SDiskbasedBuf* pBuf, SResultRowPosition* resultRowPosition, SExprSupp* pSup, SSDataBlock* pBlock, SExecTaskInfo* pTaskInfo); -SOperatorInfo* createGroupSortOperatorInfo(SOperatorInfo* downstream, SGroupSortPhysiNode* pSortPhyNode, - SExecTaskInfo* pTaskInfo); -SOperatorInfo* createTableMergeScanOperatorInfo(STableScanPhysiNode* pTableScanNode, SReadHandle* readHandle, - SExecTaskInfo* pTaskInfo); - -void copyUpdateDataBlock(SSDataBlock* pDest, SSDataBlock* pSource, int32_t tsColIndex); - bool groupbyTbname(SNodeList* pGroupList); -void* destroySqlFunctionCtx(SqlFunctionCtx* pCtx, int32_t numOfOutput); int32_t buildDataBlockFromGroupRes(SOperatorInfo* pOperator, SStreamState* pState, SSDataBlock* pBlock, SExprSupp* pSup, SGroupResInfo* pGroupResInfo); int32_t saveSessionDiscBuf(SStreamState* pState, SSessionKey* key, void* buf, int32_t size); diff --git a/source/libs/executor/src/executor.c b/source/libs/executor/src/executor.c index 9b3bd1d808..34bd9cf8ca 100644 --- a/source/libs/executor/src/executor.c +++ b/source/libs/executor/src/executor.c @@ -712,7 +712,7 @@ void qDestroyTask(qTaskInfo_t qTaskHandle) { qDebug("%s execTask completed, numOfRows:%" PRId64, GET_TASKID(pTaskInfo), pTaskInfo->pRoot->resultInfo.totalRows); - queryCostStatis(pTaskInfo); // print the query cost summary + printTaskExecCostInLog(pTaskInfo); // print the query cost summary doDestroyTask(pTaskInfo); } @@ -728,12 +728,12 @@ int32_t qSerializeTaskStatus(qTaskInfo_t tinfo, char** pOutput, int32_t* len) { } int32_t nOptrWithVal = 0; - int32_t code = encodeOperator(pTaskInfo->pRoot, pOutput, len, &nOptrWithVal); - if ((code == TSDB_CODE_SUCCESS) && (nOptrWithVal == 0)) { - taosMemoryFreeClear(*pOutput); - *len = 0; - } - return code; +// int32_t code = encodeOperator(pTaskInfo->pRoot, pOutput, len, &nOptrWithVal); +// if ((code == TSDB_CODE_SUCCESS) && (nOptrWithVal == 0)) { +// taosMemoryFreeClear(*pOutput); +// *len = 0; +// } + return 0; } int32_t qDeserializeTaskStatus(qTaskInfo_t tinfo, const char* pInput, int32_t len) { @@ -743,7 +743,8 @@ int32_t qDeserializeTaskStatus(qTaskInfo_t tinfo, const char* pInput, int32_t le return TSDB_CODE_INVALID_PARA; } - return decodeOperator(pTaskInfo->pRoot, pInput, len); + return 0; +// return decodeOperator(pTaskInfo->pRoot, pInput, len); } int32_t qExtractStreamScanner(qTaskInfo_t tinfo, void** scanner) { diff --git a/source/libs/executor/src/executorimpl.c b/source/libs/executor/src/executorimpl.c index 5eaa8ba8dd..5abde1be85 100644 --- a/source/libs/executor/src/executorimpl.c +++ b/source/libs/executor/src/executorimpl.c @@ -1335,7 +1335,7 @@ void doBuildResultDatablock(SOperatorInfo* pOperator, SOptrBasicInfo* pbInfo, SG } } -void queryCostStatis(SExecTaskInfo* pTaskInfo) { +void printTaskExecCostInLog(SExecTaskInfo* pTaskInfo) { STaskCostInfo* pSummary = &pTaskInfo->cost; SFileBlockLoadRecorder* pRecorder = pSummary->pRecoder; @@ -1958,7 +1958,7 @@ void cleanupAggSup(SAggSupporter* pAggSup) { destroyDiskbasedBuf(pAggSup->pResultBuf); } -int32_t initAggInfo(SExprSupp* pSup, SAggSupporter* pAggSup, SExprInfo* pExprInfo, int32_t numOfCols, size_t keyBufSize, +int32_t initAggSup(SExprSupp* pSup, SAggSupporter* pAggSup, SExprInfo* pExprInfo, int32_t numOfCols, size_t keyBufSize, const char* pkey) { int32_t code = initExprSupp(pSup, pExprInfo, numOfCols); if (code != TSDB_CODE_SUCCESS) { @@ -2056,7 +2056,7 @@ SOperatorInfo* createAggregateOperatorInfo(SOperatorInfo* downstream, SAggPhysiN int32_t num = 0; SExprInfo* pExprInfo = createExprInfo(pAggNode->pAggFuncs, pAggNode->pGroupKeys, &num); - int32_t code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); + int32_t code = initAggSup(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -2734,103 +2734,6 @@ int32_t rebuildReader(SOperatorInfo* pOperator, SSubplan* plan, SReadHandle* pHa } #endif -int32_t encodeOperator(SOperatorInfo* ops, char** result, int32_t* length, int32_t* nOptrWithVal) { - int32_t code = TDB_CODE_SUCCESS; - char* pCurrent = NULL; - int32_t currLength = 0; - if (ops->fpSet.encodeResultRow) { - if (result == NULL || length == NULL || nOptrWithVal == NULL) { - return TSDB_CODE_TSC_INVALID_INPUT; - } - code = ops->fpSet.encodeResultRow(ops, &pCurrent, &currLength); - - if (code != TDB_CODE_SUCCESS) { - if (*result != NULL) { - taosMemoryFree(*result); - *result = NULL; - } - return code; - } else if (currLength == 0) { - ASSERT(!pCurrent); - goto _downstream; - } - - ++(*nOptrWithVal); - - ASSERT(currLength >= 0); - - if (*result == NULL) { - *result = (char*)taosMemoryCalloc(1, currLength + sizeof(int32_t)); - if (*result == NULL) { - taosMemoryFree(pCurrent); - return TSDB_CODE_OUT_OF_MEMORY; - } - memcpy(*result + sizeof(int32_t), pCurrent, currLength); - *(int32_t*)(*result) = currLength + sizeof(int32_t); - } else { - int32_t sizePre = *(int32_t*)(*result); - char* tmp = (char*)taosMemoryRealloc(*result, sizePre + currLength); - if (tmp == NULL) { - taosMemoryFree(pCurrent); - taosMemoryFree(*result); - *result = NULL; - return TSDB_CODE_OUT_OF_MEMORY; - } - *result = tmp; - memcpy(*result + sizePre, pCurrent, currLength); - *(int32_t*)(*result) += currLength; - } - taosMemoryFree(pCurrent); - *length = *(int32_t*)(*result); - } - -_downstream: - for (int32_t i = 0; i < ops->numOfDownstream; ++i) { - code = encodeOperator(ops->pDownstream[i], result, length, nOptrWithVal); - if (code != TDB_CODE_SUCCESS) { - return code; - } - } - return TDB_CODE_SUCCESS; -} - -int32_t decodeOperator(SOperatorInfo* ops, const char* result, int32_t length) { - int32_t code = TDB_CODE_SUCCESS; - if (ops->fpSet.decodeResultRow) { - if (result == NULL) { - return TSDB_CODE_TSC_INVALID_INPUT; - } - - ASSERT(length == *(int32_t*)result); - - const char* data = result + sizeof(int32_t); - code = ops->fpSet.decodeResultRow(ops, (char*)data); - if (code != TDB_CODE_SUCCESS) { - return code; - } - - int32_t totalLength = *(int32_t*)result; - int32_t dataLength = *(int32_t*)data; - - if (totalLength == dataLength + sizeof(int32_t)) { // the last data - result = NULL; - length = 0; - } else { - result += dataLength; - *(int32_t*)(result) = totalLength - dataLength; - length = totalLength - dataLength; - } - } - - for (int32_t i = 0; i < ops->numOfDownstream; ++i) { - code = decodeOperator(ops->pDownstream[i], result, length); - if (code != TDB_CODE_SUCCESS) { - return code; - } - } - return TDB_CODE_SUCCESS; -} - int32_t createDataSinkParam(SDataSinkNode* pNode, void** pParam, qTaskInfo_t* pTaskInfo, SReadHandle* readHandle) { SExecTaskInfo* pTask = *(SExecTaskInfo**)pTaskInfo; diff --git a/source/libs/executor/src/groupoperator.c b/source/libs/executor/src/groupoperator.c index bbf9bd2a27..6dc8818900 100644 --- a/source/libs/executor/src/groupoperator.c +++ b/source/libs/executor/src/groupoperator.c @@ -456,7 +456,7 @@ SOperatorInfo* createGroupOperatorInfo(SOperatorInfo* downstream, SAggPhysiNode* int32_t num = 0; SExprInfo* pExprInfo = createExprInfo(pAggNode->pAggFuncs, pAggNode->pGroupKeys, &num); - code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, num, pInfo->groupKeyLen, pTaskInfo->id.str); + code = initAggSup(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, num, pInfo->groupKeyLen, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } diff --git a/source/libs/executor/src/projectoperator.c b/source/libs/executor/src/projectoperator.c index 819997c521..4bba3a72e1 100644 --- a/source/libs/executor/src/projectoperator.c +++ b/source/libs/executor/src/projectoperator.c @@ -102,7 +102,7 @@ SOperatorInfo* createProjectOperatorInfo(SOperatorInfo* downstream, SProjectPhys } initResultSizeInfo(&pOperator->resultInfo, numOfRows); - code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); + code = initAggSup(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -400,7 +400,7 @@ SOperatorInfo* createIndefinitOutputOperatorInfo(SOperatorInfo* downstream, SPhy initResultSizeInfo(&pOperator->resultInfo, numOfRows); blockDataEnsureCapacity(pResBlock, numOfRows); - int32_t code = initAggInfo(pSup, &pInfo->aggSup, pExprInfo, numOfExpr, keyBufSize, pTaskInfo->id.str); + int32_t code = initAggSup(pSup, &pInfo->aggSup, pExprInfo, numOfExpr, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } diff --git a/source/libs/executor/src/timewindowoperator.c b/source/libs/executor/src/timewindowoperator.c index 80c3c1c454..0e0ec5b339 100644 --- a/source/libs/executor/src/timewindowoperator.c +++ b/source/libs/executor/src/timewindowoperator.c @@ -1741,7 +1741,7 @@ SOperatorInfo* createIntervalOperatorInfo(SOperatorInfo* downstream, SIntervalPh int32_t num = 0; SExprInfo* pExprInfo = createExprInfo(pPhyNode->window.pFuncs, NULL, &num); - int32_t code = initAggInfo(pSup, &pInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); + int32_t code = initAggSup(pSup, &pInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -2001,7 +2001,7 @@ SOperatorInfo* createStatewindowOperatorInfo(SOperatorInfo* downstream, SStateWi SExprInfo* pExprInfo = createExprInfo(pStateNode->window.pFuncs, NULL, &num); initResultSizeInfo(&pOperator->resultInfo, 4096); - code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); + code = initAggSup(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -2069,7 +2069,7 @@ SOperatorInfo* createSessionAggOperatorInfo(SOperatorInfo* downstream, SSessionW SSDataBlock* pResBlock = createDataBlockFromDescNode(pSessionNode->window.node.pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); - int32_t code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); + int32_t code = initAggSup(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -2243,26 +2243,6 @@ static void clearSpecialDataBlock(SSDataBlock* pBlock) { blockDataCleanup(pBlock); } -void copyUpdateDataBlock(SSDataBlock* pDest, SSDataBlock* pSource, int32_t tsColIndex) { - // ASSERT(pDest->info.capacity >= pSource->info.rows); - blockDataEnsureCapacity(pDest, pSource->info.rows); - clearSpecialDataBlock(pDest); - SColumnInfoData* pDestCol = taosArrayGet(pDest->pDataBlock, 0); - SColumnInfoData* pSourceCol = taosArrayGet(pSource->pDataBlock, tsColIndex); - - // copy timestamp column - colDataAssign(pDestCol, pSourceCol, pSource->info.rows, &pDest->info); - for (int32_t i = 1; i < taosArrayGetSize(pDest->pDataBlock); i++) { - SColumnInfoData* pCol = taosArrayGet(pDest->pDataBlock, i); - colDataAppendNNULL(pCol, 0, pSource->info.rows); - } - - pDest->info.rows = pSource->info.rows; - pDest->info.groupId = pSource->info.groupId; - pDest->info.type = pSource->info.type; - blockDataUpdateTsWindow(pDest, 0); -} - static void doBuildPullDataBlock(SArray* array, int32_t* pIndex, SSDataBlock* pBlock) { clearSpecialDataBlock(pBlock); int32_t size = taosArrayGetSize(array); @@ -2707,7 +2687,7 @@ SOperatorInfo* createStreamFinalIntervalOperatorInfo(SOperatorInfo* downstream, SSDataBlock* pResBlock = createDataBlockFromDescNode(pPhyNode->pOutputDataBlockDesc); initBasicInfo(&pInfo->binfo, pResBlock); - int32_t code = initAggInfo(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); + int32_t code = initAggSup(&pOperator->exprSupp, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -4346,7 +4326,7 @@ SOperatorInfo* createMergeAlignedIntervalOperatorInfo(SOperatorInfo* downstream, int32_t num = 0; SExprInfo* pExprInfo = createExprInfo(pNode->window.pFuncs, NULL, &num); - code = initAggInfo(&pOperator->exprSupp, &iaInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); + code = initAggSup(&pOperator->exprSupp, &iaInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -4651,7 +4631,7 @@ SOperatorInfo* createMergeIntervalOperatorInfo(SOperatorInfo* downstream, SMerge size_t keyBufSize = sizeof(int64_t) + sizeof(int64_t) + POINTER_BYTES; initResultSizeInfo(&pOperator->resultInfo, 4096); - int32_t code = initAggInfo(pExprSupp, &pIntervalInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); + int32_t code = initAggSup(pExprSupp, &pIntervalInfo->aggSup, pExprInfo, num, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -4847,7 +4827,7 @@ SOperatorInfo* createStreamIntervalOperatorInfo(SOperatorInfo* downstream, SPhys initResultSizeInfo(&pOperator->resultInfo, 4096); size_t keyBufSize = sizeof(int64_t) + sizeof(int64_t) + POINTER_BYTES; - code = initAggInfo(pSup, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); + code = initAggSup(pSup, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str); if (code != TSDB_CODE_SUCCESS) { goto _error; } @@ -4901,3 +4881,4 @@ _error: pTaskInfo->code = code; return NULL; } +