Merge pull request #12651 from taosdata/feature/3.0_glzhao

enh(query): support sum(NULL), max(NULL), min(NULL), count(NULL)

include/common/ttypes.h

@@ -179,6 +179,8 @@ typedef struct {
     }                                          \
   } while (0)
 
+//TODO: use varchar(0) to represent NULL type
+#define IS_NULL_TYPE(_t)             ((_t) == TSDB_DATA_TYPE_NULL)
 #define IS_SIGNED_NUMERIC_TYPE(_t)   ((_t) >= TSDB_DATA_TYPE_TINYINT && (_t) <= TSDB_DATA_TYPE_BIGINT)
 #define IS_UNSIGNED_NUMERIC_TYPE(_t) ((_t) >= TSDB_DATA_TYPE_UTINYINT && (_t) <= TSDB_DATA_TYPE_UBIGINT)
 #define IS_FLOAT_TYPE(_t)            ((_t) == TSDB_DATA_TYPE_FLOAT || (_t) == TSDB_DATA_TYPE_DOUBLE)

source/libs/function/inc/builtinsimpl.h

@@ -74,6 +74,7 @@ int32_t diffFunction(SqlFunctionCtx *pCtx);
 bool getFirstLastFuncEnv(struct SFunctionNode* pFunc, SFuncExecEnv* pEnv);
 int32_t firstFunction(SqlFunctionCtx *pCtx);
 int32_t lastFunction(SqlFunctionCtx *pCtx);
+int32_t lastFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock);
 
 bool getTopBotFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv);
 int32_t topFunction(SqlFunctionCtx *pCtx);

source/libs/function/src/builtins.c

@@ -47,8 +47,10 @@ static int32_t translateInOutNum(SFunctionNode* pFunc, char* pErrBuf, int32_t le
   }
 
   uint8_t paraType = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 0))->resType.type;
-  if (!IS_NUMERIC_TYPE(paraType)) {
+  if (!IS_NUMERIC_TYPE(paraType) && !IS_NULL_TYPE(paraType)) {
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
+  } else if (IS_NULL_TYPE(paraType)) {
+    paraType = TSDB_DATA_TYPE_BIGINT;
   }
 
   pFunc->node.resType = (SDataType){.bytes = tDataTypes[paraType].bytes, .type = paraType};
@@ -62,7 +64,7 @@ static int32_t translateInNumOutDou(SFunctionNode* pFunc, char* pErrBuf, int32_t
   }
 
   uint8_t paraType = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 0))->resType.type;
-  if (!IS_NUMERIC_TYPE(paraType)) {
+  if (!IS_NUMERIC_TYPE(paraType) && !IS_NULL_TYPE(paraType)) {
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
@@ -115,18 +117,19 @@ static int32_t translateSum(SFunctionNode* pFunc, char* pErrBuf, int32_t len) {
   }
 
   uint8_t paraType = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 0))->resType.type;
-  if (!IS_NUMERIC_TYPE(paraType)) {
+  if (!IS_NUMERIC_TYPE(paraType) && !IS_NULL_TYPE(paraType)) {
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
   uint8_t resType = 0;
-  if (IS_SIGNED_NUMERIC_TYPE(paraType) || paraType == TSDB_DATA_TYPE_BOOL) {
+  if (IS_SIGNED_NUMERIC_TYPE(paraType) || TSDB_DATA_TYPE_BOOL == paraType || IS_NULL_TYPE(paraType)) {
     resType = TSDB_DATA_TYPE_BIGINT;
   } else if (IS_UNSIGNED_NUMERIC_TYPE(paraType)) {
     resType = TSDB_DATA_TYPE_UBIGINT;
   } else if (IS_FLOAT_TYPE(paraType)) {
     resType = TSDB_DATA_TYPE_DOUBLE;
   }
+
   pFunc->node.resType = (SDataType){.bytes = tDataTypes[resType].bytes, .type = resType};
   return TSDB_CODE_SUCCESS;
 }
@@ -872,7 +875,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
     .getEnvFunc   = getFirstLastFuncEnv,
     .initFunc     = functionSetup,
     .processFunc  = lastFunction,
-    .finalizeFunc = functionFinalize
+    .finalizeFunc = lastFinalize
   },
   {
     .name = "diff",

source/libs/function/src/builtinsimpl.c

@@ -255,7 +255,7 @@ int32_t functionFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
   SColumnInfoData* pCol = taosArrayGet(pBlock->pDataBlock, slotId);
 
   SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx);
-  pResInfo->isNullRes = (pResInfo->numOfRes == 0) ? 1 : 0;
+  //pResInfo->isNullRes = (pResInfo->numOfRes == 0) ? 1 : 0;
 
   char* in = GET_ROWCELL_INTERBUF(pResInfo);
   colDataAppend(pCol, pBlock->info.rows, in, pResInfo->isNullRes);
@@ -331,8 +331,18 @@ static FORCE_INLINE int32_t getNumofElem(SqlFunctionCtx* pCtx) {
 int32_t countFunction(SqlFunctionCtx* pCtx) {
   int32_t numOfElem = getNumofElem(pCtx);
   SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx);
+
+  SInputColumnInfoData* pInput = &pCtx->input;
+  int32_t type   = pInput->pData[0]->info.type;
+
   char* buf = GET_ROWCELL_INTERBUF(pResInfo);
+  if (IS_NULL_TYPE(type)) {
+    //select count(NULL) returns 0
+    numOfElem = 1;
+    *((int64_t*)buf) = 0;
+  } else {
     *((int64_t*)buf) += numOfElem;
+  }
 
   SET_VAL(pResInfo, numOfElem, 1);
   return TSDB_CODE_SUCCESS;
@@ -383,6 +393,12 @@ int32_t sumFunction(SqlFunctionCtx* pCtx) {
 
   SSumRes* pSumRes = GET_ROWCELL_INTERBUF(GET_RES_INFO(pCtx));
 
+  if (IS_NULL_TYPE(type)) {
+    GET_RES_INFO(pCtx)->isNullRes = 1;
+    numOfElem = 1;
+    goto _sum_over;
+  }
+
   if (pInput->colDataAggIsSet) {
     numOfElem = pInput->numOfRows - pAgg->numOfNull;
     ASSERT(numOfElem >= 0);
@@ -427,6 +443,7 @@ int32_t sumFunction(SqlFunctionCtx* pCtx) {
     }
   }
 
+_sum_over:
   // data in the check operation are all null, not output
   SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1);
   return TSDB_CODE_SUCCESS;
@@ -526,6 +543,12 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) {
   int32_t start = pInput->startRowIndex;
   int32_t numOfRows = pInput->numOfRows;
 
+  if (IS_NULL_TYPE(type)) {
+    GET_RES_INFO(pCtx)->isNullRes = 1;
+    numOfElem = 1;
+    goto _avg_over;
+  }
+
   switch (type) {
     case TSDB_DATA_TYPE_TINYINT: {
       int8_t* plist = (int8_t*)pCol->pData;
@@ -617,6 +640,7 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) {
       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;
@@ -783,6 +807,12 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) {
   SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx);
   SMinmaxResInfo *pBuf = GET_ROWCELL_INTERBUF(pResInfo);
 
+  if (IS_NULL_TYPE(type)) {
+    GET_RES_INFO(pCtx)->isNullRes = 1;
+    numOfElems = 1;
+    goto _min_max_over;
+  }
+
   // data in current data block are qualified to the query
   if (pInput->colDataAggIsSet) {
     numOfElems = pInput->numOfRows - pAgg->numOfNull;
@@ -1183,6 +1213,7 @@ int32_t doMinMaxHelper(SqlFunctionCtx* pCtx, int32_t isMinFunc) {
     }
   }
 
+_min_max_over:
   return numOfElems;
 }
 
@@ -1215,9 +1246,9 @@ int32_t minmaxFunctionFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
 
   if (pCol->info.type == TSDB_DATA_TYPE_FLOAT) {
     float v = *(double*) &pRes->v;
-    colDataAppend(pCol, currentRow, (const char*)&v, false);
+    colDataAppend(pCol, currentRow, (const char*)&v, pEntryInfo->isNullRes);
   } else {
-    colDataAppend(pCol, currentRow, (const char*)&pRes->v, false);
+    colDataAppend(pCol, currentRow, (const char*)&pRes->v, pEntryInfo->isNullRes);
   }
 
   setSelectivityValue(pCtx, pBlock, &pRes->tuplePos, currentRow);
@@ -1287,6 +1318,12 @@ int32_t stddevFunction(SqlFunctionCtx* pCtx) {
   int32_t start = pInput->startRowIndex;
   int32_t numOfRows = pInput->numOfRows;
 
+  if (IS_NULL_TYPE(type)) {
+    GET_RES_INFO(pCtx)->isNullRes = 1;
+    numOfElem = 1;
+    goto _stddev_over;
+  }
+
   switch (type) {
     case TSDB_DATA_TYPE_TINYINT: {
       int8_t* plist = (int8_t*)pCol->pData;
@@ -1384,6 +1421,7 @@ int32_t stddevFunction(SqlFunctionCtx* pCtx) {
       break;
   }
 
+_stddev_over:
   // data in the check operation are all null, not output
   SET_VAL(GET_RES_INFO(pCtx), numOfElem, 1);
   return TSDB_CODE_SUCCESS;
@@ -1943,6 +1981,19 @@ int32_t lastFunction(SqlFunctionCtx* pCtx) {
   return TSDB_CODE_SUCCESS;
 }
 
+int32_t lastFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
+  int32_t          slotId = pCtx->pExpr->base.resSchema.slotId;
+  SColumnInfoData* pCol = taosArrayGet(pBlock->pDataBlock, slotId);
+
+  SResultRowEntryInfo* pResInfo = GET_RES_INFO(pCtx);
+  pResInfo->isNullRes = (pResInfo->numOfRes == 0) ? 1 : 0;
+
+  char* in = GET_ROWCELL_INTERBUF(pResInfo);
+  colDataAppend(pCol, pBlock->info.rows, in, pResInfo->isNullRes);
+
+  return pResInfo->numOfRes;
+}
+
 bool getDiffFuncEnv(SFunctionNode* UNUSED_PARAM(pFunc), SFuncExecEnv* pEnv) {
   pEnv->calcMemSize = sizeof(SDiffInfo);
   return true;