enh: optimize group by tag queries

This commit is contained in:
dapan1121 2025-02-17 16:18:13 +08:00
parent c0aa91ff1e
commit 12dd3a621e
4 changed files with 64 additions and 1 deletions

View File

@ -128,6 +128,7 @@ typedef struct SScanLogicNode {
bool paraTablesSort; // for table merge scan
bool smallDataTsSort; // disable row id sort for table merge scan
bool needSplit;
bool noPseudoRefAfterGrp; // no pseudo columns referenced ater group/partition clause
} SScanLogicNode;
typedef struct SJoinLogicNode {

View File

@ -510,6 +510,7 @@ static int32_t logicScanCopy(const SScanLogicNode* pSrc, SScanLogicNode* pDst) {
COPY_SCALAR_FIELD(paraTablesSort);
COPY_SCALAR_FIELD(smallDataTsSort);
COPY_SCALAR_FIELD(needSplit);
COPY_SCALAR_FIELD(noPseudoRefAfterGrp);
return TSDB_CODE_SUCCESS;
}

View File

@ -406,6 +406,47 @@ static int32_t makeScanLogicNode(SLogicPlanContext* pCxt, SRealTableNode* pRealT
static bool needScanDefaultCol(EScanType scanType) { return SCAN_TYPE_TABLE_COUNT != scanType; }
static int32_t updateScanNoPseudoRefAfterGrp(SSelectStmt* pSelect, SScanLogicNode* pScan, SRealTableNode* pRealTable) {
if (NULL == pScan->pScanPseudoCols || pScan->pScanPseudoCols->length <= 0) {
return TSDB_CODE_SUCCESS;
}
SNodeList* pList = NULL;
int32_t code = 0;
if (NULL == pSelect->pPartitionByList || pSelect->pPartitionByList->length <= 0) {
if (NULL == pSelect->pGroupByList || pSelect->pGroupByList->length <= 0) {
return TSDB_CODE_SUCCESS;
}
code = nodesCollectColumns(pSelect, SQL_CLAUSE_GROUP_BY, pRealTable->table.tableAlias, COLLECT_COL_TYPE_TAG,
&pList);
if (TSDB_CODE_SUCCESS == code) {
code = nodesCollectFuncs(pSelect, SQL_CLAUSE_GROUP_BY, pRealTable->table.tableAlias, fmIsScanPseudoColumnFunc,
&pList);
}
if (TSDB_CODE_SUCCESS == code && (NULL == pList || pList->length <= 0)) {
pScan->noPseudoRefAfterGrp = true;
}
goto _return;
}
code = nodesCollectColumns(pSelect, SQL_CLAUSE_PARTITION_BY, pRealTable->table.tableAlias, COLLECT_COL_TYPE_TAG,
&pList);
if (TSDB_CODE_SUCCESS == code) {
code = nodesCollectFuncs(pSelect, SQL_CLAUSE_PARTITION_BY, pRealTable->table.tableAlias, fmIsScanPseudoColumnFunc,
&pList);
}
if (TSDB_CODE_SUCCESS == code && (NULL == pList || pList->length <= 0)) {
pScan->noPseudoRefAfterGrp = true;
}
_return:
nodesDestroyList(pList);
return code;
}
static int32_t createScanLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect, SRealTableNode* pRealTable,
SLogicNode** pLogicNode) {
SScanLogicNode* pScan = NULL;
@ -437,7 +478,13 @@ static int32_t createScanLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect
&pScan->pScanPseudoCols);
}
pScan->scanType = getScanType(pCxt, pScan->pScanPseudoCols, pScan->pScanCols, pScan->tableType, pSelect->tagScan);
if (TSDB_CODE_SUCCESS == code) {
code = updateScanNoPseudoRefAfterGrp(pSelect, pScan, pRealTable);
}
if (TSDB_CODE_SUCCESS == code) {
pScan->scanType = getScanType(pCxt, pScan->pScanPseudoCols, pScan->pScanCols, pScan->tableType, pSelect->tagScan);
}
// rewrite the expression in subsequent clauses
if (TSDB_CODE_SUCCESS == code) {

View File

@ -3302,6 +3302,17 @@ static int32_t partTagsRewriteGroupTagsToFuncs(SNodeList* pGroupTags, int32_t st
return code;
}
static int32_t partTagsOptRemovePseudoCols(SScanLogicNode* pScan) {
if (!pScan->noPseudoRefAfterGrp || NULL == pScan->pScanPseudoCols || pScan->pScanPseudoCols->length <= 0) {
return TSDB_CODE_SUCCESS;
}
nodesDestroyList(pScan->pScanPseudoCols);
pScan->pScanPseudoCols = NULL;
return TSDB_CODE_SUCCESS;
}
static int32_t partTagsOptimize(SOptimizeContext* pCxt, SLogicSubplan* pLogicSubplan) {
SLogicNode* pNode = optFindPossibleNode(pLogicSubplan->pNode, partTagsOptMayBeOptimized, NULL);
if (NULL == pNode) {
@ -3362,6 +3373,9 @@ static int32_t partTagsOptimize(SOptimizeContext* pCxt, SLogicSubplan* pLogicSub
code = partTagsRewriteGroupTagsToFuncs(pScan->pGroupTags, start, pAgg);
}
}
if (TSDB_CODE_SUCCESS == code) {
code = partTagsOptRemovePseudoCols(pScan);
}
if (TSDB_CODE_SUCCESS == code) {
code = partTagsOptRebuildTbanme(pScan->pGroupTags);
}