statistics, executor: merge Count and ModifyCount for global stats

executor/analyze_global_stats.go

@@ -83,7 +83,17 @@ func (e *AnalyzeExec) handleGlobalStats(ctx context.Context, needGlobalStats boo
 				for i := 0; i < globalStats.Num; i++ {
 					hg, cms, topN := globalStats.Hg[i], globalStats.Cms[i], globalStats.TopN[i]
 					// fms for global stats doesn't need to dump to kv.
-					err = statsHandle.SaveStatsToStorage(globalStatsID.tableID, globalStats.Count, info.isIndex, hg, cms, topN, info.statsVersion, 1, true)
+					err = statsHandle.SaveStatsToStorage(globalStatsID.tableID,
+						globalStats.Count,
+						globalStats.ModifyCount,
+						info.isIndex,
+						hg,
+						cms,
+						topN,
+						info.statsVersion,
+						1,
+						true,
+					)
 					if err != nil {
 						logutil.Logger(ctx).Error("save global-level stats to storage failed", zap.Error(err))
 					}

statistics/handle/ddl.go

@@ -156,7 +156,7 @@ func (h *Handle) updateGlobalStats(tblInfo *model.TableInfo) error {
 	for i := 0; i < newColGlobalStats.Num; i++ {
 		hg, cms, topN := newColGlobalStats.Hg[i], newColGlobalStats.Cms[i], newColGlobalStats.TopN[i]
 		// fms for global stats doesn't need to dump to kv.
-		err = h.SaveStatsToStorage(tableID, newColGlobalStats.Count, 0, hg, cms, topN, 2, 1, false)
+		err = h.SaveStatsToStorage(tableID, newColGlobalStats.Count, newColGlobalStats.ModifyCount, 0, hg, cms, topN, 2, 1, false)
 		if err != nil {
 			return err
 		}
@@ -186,7 +186,7 @@ func (h *Handle) updateGlobalStats(tblInfo *model.TableInfo) error {
 		for i := 0; i < newIndexGlobalStats.Num; i++ {
 			hg, cms, topN := newIndexGlobalStats.Hg[i], newIndexGlobalStats.Cms[i], newIndexGlobalStats.TopN[i]
 			// fms for global stats doesn't need to dump to kv.
-			err = h.SaveStatsToStorage(tableID, newIndexGlobalStats.Count, 1, hg, cms, topN, 2, 1, false)
+			err = h.SaveStatsToStorage(tableID, newIndexGlobalStats.Count, newIndexGlobalStats.ModifyCount, 1, hg, cms, topN, 2, 1, false)
 			if err != nil {
 				return err
 			}

statistics/handle/dump.go

@@ -252,14 +252,18 @@ func (h *Handle) loadStatsFromJSON(tableInfo *model.TableInfo, physicalID int64,
 
 	for _, col := range tbl.Columns {
 		// loadStatsFromJSON doesn't support partition table now.
-		err = h.SaveStatsToStorage(tbl.PhysicalID, tbl.Count, 0, &col.Histogram, col.CMSketch, col.TopN, int(col.StatsVer), 1, false)
+		// The table level Count and Modify_count would be overridden by the SaveMetaToStorage below, so we don't need
+		// to care about them here.
+		err = h.SaveStatsToStorage(tbl.PhysicalID, tbl.Count, 0, 0, &col.Histogram, col.CMSketch, col.TopN, int(col.StatsVer), 1, false)
 		if err != nil {
 			return errors.Trace(err)
 		}
 	}
 	for _, idx := range tbl.Indices {
 		// loadStatsFromJSON doesn't support partition table now.
-		err = h.SaveStatsToStorage(tbl.PhysicalID, tbl.Count, 1, &idx.Histogram, idx.CMSketch, idx.TopN, int(idx.StatsVer), 1, false)
+		// The table level Count and Modify_count would be overridden by the SaveMetaToStorage below, so we don't need
+		// to care about them here.
+		err = h.SaveStatsToStorage(tbl.PhysicalID, tbl.Count, 0, 1, &idx.Histogram, idx.CMSketch, idx.TopN, int(idx.StatsVer), 1, false)
 		if err != nil {
 			return errors.Trace(err)
 		}

statistics/handle/dump_test.go

@@ -261,7 +261,7 @@ func TestDumpCMSketchWithTopN(t *testing.T) {
 	cms, _, _, _ := statistics.NewCMSketchAndTopN(5, 2048, fakeData, 20, 100)
 
 	stat := h.GetTableStats(tableInfo)
-	err = h.SaveStatsToStorage(tableInfo.ID, 1, 0, &stat.Columns[tableInfo.Columns[0].ID].Histogram, cms, nil, statistics.Version2, 1, false)
+	err = h.SaveStatsToStorage(tableInfo.ID, 1, 0, 0, &stat.Columns[tableInfo.Columns[0].ID].Histogram, cms, nil, statistics.Version2, 1, false)
 	require.NoError(t, err)
 	require.Nil(t, h.Update(is))
 

statistics/handle/handle.go

@@ -399,12 +399,13 @@ func (h *Handle) UpdateSessionVar() error {
 // In the column statistics, the variable `num` is equal to the number of columns in the partition table.
 // In the index statistics, the variable `num` is always equal to one.
 type GlobalStats struct {
-	Num   int
-	Count int64
-	Hg    []*statistics.Histogram
-	Cms   []*statistics.CMSketch
-	TopN  []*statistics.TopN
-	Fms   []*statistics.FMSketch
+	Num         int
+	Count       int64
+	ModifyCount int64
+	Hg          []*statistics.Histogram
+	Cms         []*statistics.CMSketch
+	TopN        []*statistics.TopN
+	Fms         []*statistics.FMSketch
 }
 
 // MergePartitionStats2GlobalStatsByTableID merge the partition-level stats to global-level stats based on the tableID.
@@ -513,7 +514,7 @@ func (h *Handle) mergePartitionStats2GlobalStats(sc sessionctx.Context,
 			allPartitionStats[partitionID] = partitionStats
 		}
 		for i := 0; i < globalStats.Num; i++ {
-			count, hg, cms, topN, fms := partitionStats.GetStatsInfo(histIDs[i], isIndex == 1)
+			_, hg, cms, topN, fms := partitionStats.GetStatsInfo(histIDs[i], isIndex == 1)
 			// partition stats is not empty but column stats(hist, topn) is missing
 			if partitionStats.Count > 0 && (hg == nil || hg.TotalRowCount() <= 0) && (topN == nil || topN.TotalCount() <= 0) {
 				var errMsg string
@@ -527,7 +528,8 @@ func (h *Handle) mergePartitionStats2GlobalStats(sc sessionctx.Context,
 			}
 			if i == 0 {
 				// In a partition, we will only update globalStats.Count once
-				globalStats.Count += count
+				globalStats.Count += partitionStats.Count
+				globalStats.ModifyCount += partitionStats.ModifyCount
 			}
 			allHg[i] = append(allHg[i], hg)
 			allCms[i] = append(allCms[i], cms)
@@ -1532,8 +1534,10 @@ func SaveTableStatsToStorage(sctx sessionctx.Context, results *statistics.Analyz
 }
 
 // SaveStatsToStorage saves the stats to storage.
+// If count is negative, both count and modify count would not be used and not be written to the table. Unless, corresponding
+// fields in the stats_meta table will be updated.
 // TODO: refactor to reduce the number of parameters
-func (h *Handle) SaveStatsToStorage(tableID int64, count int64, isIndex int, hg *statistics.Histogram, cms *statistics.CMSketch, topN *statistics.TopN, statsVersion int, isAnalyzed int64, updateAnalyzeTime bool) (err error) {
+func (h *Handle) SaveStatsToStorage(tableID int64, count, modifyCount int64, isIndex int, hg *statistics.Histogram, cms *statistics.CMSketch, topN *statistics.TopN, statsVersion int, isAnalyzed int64, updateAnalyzeTime bool) (err error) {
 	statsVer := uint64(0)
 	defer func() {
 		if err == nil && statsVer != 0 {
@@ -1559,7 +1563,7 @@ func (h *Handle) SaveStatsToStorage(tableID int64, count int64, isIndex int, hg
 	version := txn.StartTS()
 	// If the count is less than 0, then we do not want to update the modify count and count.
 	if count >= 0 {
-		_, err = exec.ExecuteInternal(ctx, "replace into mysql.stats_meta (version, table_id, count) values (%?, %?, %?)", version, tableID, count)
+		_, err = exec.ExecuteInternal(ctx, "replace into mysql.stats_meta (version, table_id, count, modify_count) values (%?, %?, %?, %?)", version, tableID, count, modifyCount)
 	} else {
 		_, err = exec.ExecuteInternal(ctx, "update mysql.stats_meta set version = %? where table_id = %?", version, tableID)
 	}

statistics/handle/handle_test.go

@@ -1659,6 +1659,7 @@ partition by range (a) (
 	partition p0 values less than (10),
 	partition p1 values less than (20)
 )`)
+	require.NoError(t, dom.StatsHandle().HandleDDLEvent(<-dom.StatsHandle().DDLEventCh()))
 	tk.MustExec("insert into t values (1), (5), (null), (11), (15)")
 	require.NoError(t, dom.StatsHandle().DumpStatsDeltaToKV(handle.DumpAll))
 
@@ -1694,14 +1695,15 @@ partition by range (a) (
 	require.NoError(t, err)
 	tableInfo := tbl.Meta()
 	globalStats := h.GetTableStats(tableInfo)
-	// global.count = p0.count(3) + p1.count(2) + p2.count(2)
-	// We did not analyze partition p1, so the value here has not changed
-	require.Equal(t, int64(7), globalStats.Count)
+	// global.count = p0.count(3) + p1.count(4) + p2.count(2)
+	// modify count is 2 because we didn't analyze p1 after the second insert
+	require.Equal(t, int64(9), globalStats.Count)
+	require.Equal(t, int64(2), globalStats.ModifyCount)
 
 	tk.MustExec("analyze table t partition p1;")
 	globalStats = h.GetTableStats(tableInfo)
 	// global.count = p0.count(3) + p1.count(4) + p2.count(4)
-	// The value of p1.Count is correct now.
+	// The value of modify count is 0 now.
 	require.Equal(t, int64(9), globalStats.Count)
 	require.Equal(t, int64(0), globalStats.ModifyCount)
 

statistics/handle/update.go

@@ -886,7 +886,7 @@ func (h *Handle) deleteOutdatedFeedback(tableID, histID, isIndex int64) error {
 func (h *Handle) dumpStatsUpdateToKV(tableID, isIndex int64, q *statistics.QueryFeedback, hist *statistics.Histogram, cms *statistics.CMSketch, topN *statistics.TopN, statsVersion int64) error {
 	hist = statistics.UpdateHistogram(hist, q, int(statsVersion))
 	// feedback for partition is not ready.
-	err := h.SaveStatsToStorage(tableID, -1, int(isIndex), hist, cms, topN, int(statsVersion), 0, false)
+	err := h.SaveStatsToStorage(tableID, -1, 0, int(isIndex), hist, cms, topN, int(statsVersion), 0, false)
 	metrics.UpdateStatsCounter.WithLabelValues(metrics.RetLabel(err)).Inc()
 	return errors.Trace(err)
 }