planner: add tolerance for row count floating point comparison (#59241) (#59278)

close #59133

Author: ti-chi-bot

pkg/planner/core/cost/factors_thresholds.go

@@ -24,6 +24,10 @@ const (
 	SelectionFactor = 0.8
 
 	DistinctFactor = 0.8
+
+	// ToleranceFactor is an arbitrary value used in (some) floating point
+	// comparisons to account for precision errors
+	ToleranceFactor = 0.00001
 )
 
 // AggFuncFactor is the basic factor for aggregation.

pkg/planner/core/find_best_task.go

@@ -1680,7 +1680,8 @@ func convertToIndexMergeScan(ds *logicalop.DataSource, prop *property.PhysicalPr
 		scans = append(scans, scan)
 	}
 	totalRowCount := path.CountAfterAccess
-	if prop.ExpectedCnt < ds.StatsInfo().RowCount {
+	// Add an arbitrary tolerance factor to account for comparison with floating point
+	if (prop.ExpectedCnt + cost.ToleranceFactor) < ds.StatsInfo().RowCount {
 		totalRowCount *= prop.ExpectedCnt / ds.StatsInfo().RowCount
 	}
 	ts, remainingFilters2, moreColumn, err := buildIndexMergeTableScan(ds, path.TableFilters, totalRowCount, candidate.isMatchProp)
@@ -2925,7 +2926,8 @@ func getOriginalPhysicalTableScan(ds *logicalop.DataSource, prop *property.Physi
 	}.Init(ds.SCtx(), ds.QueryBlockOffset())
 	ts.SetSchema(ds.Schema().Clone())
 	rowCount := path.CountAfterAccess
-	if prop.ExpectedCnt < ds.StatsInfo().RowCount {
+	// Add an arbitrary tolerance factor to account for comparison with floating point
+	if (prop.ExpectedCnt + cost.ToleranceFactor) < ds.StatsInfo().RowCount {
 		rowCount = cardinality.AdjustRowCountForTableScanByLimit(ds.SCtx(),
 			ds.StatsInfo(), ds.TableStats, ds.StatisticTable,
 			path, prop.ExpectedCnt, isMatchProp && prop.SortItems[0].Desc)

pkg/planner/core/stats.go

@@ -234,7 +234,8 @@ func deriveIndexPathStats(ds *logicalop.DataSource, path *util.AccessPath, _ []e
 	path.IndexFilters = append(path.IndexFilters, indexFilters...)
 	// If the `CountAfterAccess` is less than `stats.RowCount`, there must be some inconsistent stats info.
 	// We prefer the `stats.RowCount` because it could use more stats info to calculate the selectivity.
-	if path.CountAfterAccess < ds.StatsInfo().RowCount && !isIm {
+	// Add an arbitrary tolerance factor to account for comparison with floating point
+	if (path.CountAfterAccess+cost.ToleranceFactor) < ds.StatsInfo().RowCount && !isIm {
 		path.CountAfterAccess = math.Min(ds.StatsInfo().RowCount/cost.SelectionFactor, float64(ds.StatisticTable.RealtimeCount))
 	}
 	if path.IndexFilters != nil {
@@ -333,7 +334,8 @@ func deriveTablePathStats(ds *logicalop.DataSource, path *util.AccessPath, conds
 	path.CountAfterAccess, err = cardinality.GetRowCountByIntColumnRanges(ds.SCtx(), &ds.StatisticTable.HistColl, pkCol.ID, path.Ranges)
 	// If the `CountAfterAccess` is less than `stats.RowCount`, there must be some inconsistent stats info.
 	// We prefer the `stats.RowCount` because it could use more stats info to calculate the selectivity.
-	if path.CountAfterAccess < ds.StatsInfo().RowCount && !isIm {
+	// Add an arbitrary tolerance factor to account for comparison with floating point
+	if (path.CountAfterAccess+cost.ToleranceFactor) < ds.StatsInfo().RowCount && !isIm {
 		path.CountAfterAccess = math.Min(ds.StatsInfo().RowCount/cost.SelectionFactor, float64(ds.StatisticTable.RealtimeCount))
 	}
 	return err
@@ -371,7 +373,8 @@ func deriveCommonHandleTablePathStats(ds *logicalop.DataSource, path *util.Acces
 	}
 	// If the `CountAfterAccess` is less than `stats.RowCount`, there must be some inconsistent stats info.
 	// We prefer the `stats.RowCount` because it could use more stats info to calculate the selectivity.
-	if path.CountAfterAccess < ds.StatsInfo().RowCount && !isIm {
+	// Add an arbitrary tolerance factor to account for comparison with floating point
+	if (path.CountAfterAccess+cost.ToleranceFactor) < ds.StatsInfo().RowCount && !isIm {
 		path.CountAfterAccess = math.Min(ds.StatsInfo().RowCount/cost.SelectionFactor, float64(ds.StatisticTable.RealtimeCount))
 	}
 	return nil

tests/integrationtest/r/clustered_index.result

@@ -109,11 +109,11 @@ id	estRows	task	access object	operator info
 StreamAgg_17	1.00	root		funcs:count(Column#8)->Column#6
 └─IndexReader_18	1.00	root		index:StreamAgg_9
   └─StreamAgg_9	1.00	cop[tikv]		funcs:count(1)->Column#8
-    └─IndexRangeScan_16	133.89	cop[tikv]	table:tbl_0, index:idx_3(col_0)	range:[803163,+inf], keep order:false
+    └─IndexRangeScan_16	107.12	cop[tikv]	table:tbl_0, index:idx_3(col_0)	range:[803163,+inf], keep order:false
 explain select count(*) from wout_cluster_index.tbl_0 where col_0 >= 803163  ;
 id	estRows	task	access object	operator info
 StreamAgg_17	1.00	root		funcs:count(Column#9)->Column#7
 └─IndexReader_18	1.00	root		index:StreamAgg_9
   └─StreamAgg_9	1.00	cop[tikv]		funcs:count(1)->Column#9
-    └─IndexRangeScan_16	133.89	cop[tikv]	table:tbl_0, index:idx_3(col_0)	range:[803163,+inf], keep order:false
+    └─IndexRangeScan_16	107.12	cop[tikv]	table:tbl_0, index:idx_3(col_0)	range:[803163,+inf], keep order:false
 set @@tidb_enable_outer_join_reorder=false;