statistics: correct behavior of non-lite InitStats and stats sync load of no stats column (#57803) (#59590)

close #57804

Author: ti-chi-bot

pkg/statistics/handle/BUILD.bazel

@@ -59,7 +59,7 @@ go_test(
     embed = [":handle"],
     flaky = True,
     race = "on",
-    shard_count = 11,
+    shard_count = 12,
     deps = [
         "//pkg/config",
         "//pkg/parser/model",

pkg/statistics/handle/bootstrap.go

@@ -117,6 +117,16 @@ func (h *Handle) initStatsMeta(is infoschema.InfoSchema) (util.StatsCache, error
 	return tables, nil
 }
 
+// initStatsHistogramsSQLGen generates the SQL to load all stats_histograms records.
+func initStatsHistogramsSQLGen(isPaging bool) string {
+	selectPrefix := "select /*+ ORDER_INDEX(mysql.stats_histograms,tbl) */ HIGH_PRIORITY table_id, is_index, hist_id, distinct_count, version, null_count, cm_sketch, tot_col_size, stats_ver, correlation, flag, last_analyze_pos from mysql.stats_histograms"
+	orderSuffix := " order by table_id"
+	if !isPaging {
+		return selectPrefix + orderSuffix
+	}
+	return selectPrefix + " where table_id >= %? and table_id < %?" + orderSuffix
+}
+
 func (h *Handle) initStatsHistograms4ChunkLite(is infoschema.InfoSchema, cache util.StatsCache, iter *chunk.Iterator4Chunk) {
 	var table *statistics.Table
 	for row := iter.Begin(); row != iter.End(); row = iter.Next() {
@@ -137,9 +147,9 @@ func (h *Handle) initStatsHistograms4ChunkLite(is infoschema.InfoSchema, cache u
 		ndv := row.GetInt64(3)
 		version := row.GetUint64(4)
 		nullCount := row.GetInt64(5)
-		statsVer := row.GetInt64(7)
-		flag := row.GetInt64(9)
-		lastAnalyzePos := row.GetDatum(10, types.NewFieldType(mysql.TypeBlob))
+		statsVer := row.GetInt64(8)
+		flag := row.GetInt64(10)
+		lastAnalyzePos := row.GetDatum(11, types.NewFieldType(mysql.TypeBlob))
 		tbl, _ := h.TableInfoByID(is, table.PhysicalID)
 		if isIndex > 0 {
 			var idxInfo *model.IndexInfo
@@ -176,7 +186,7 @@ func (h *Handle) initStatsHistograms4ChunkLite(is infoschema.InfoSchema, cache u
 			if colInfo == nil {
 				continue
 			}
-			hist := statistics.NewHistogram(id, ndv, nullCount, version, &colInfo.FieldType, 0, row.GetInt64(6))
+			hist := statistics.NewHistogram(id, ndv, nullCount, version, &colInfo.FieldType, 0, row.GetInt64(7))
 			hist.Correlation = row.GetFloat64(8)
 			col := &statistics.Column{
 				Histogram:  *hist,
@@ -288,7 +298,9 @@ func (h *Handle) initStatsHistograms4Chunk(is infoschema.InfoSchema, cache util.
 				StatsVer:   statsVer,
 			}
 			// primary key column has no stats info, because primary key's is_index is false. so it cannot load the topn
-			col.StatsLoadedStatus = statistics.NewStatsAllEvictedStatus()
+			if col.StatsAvailable() {
+				col.StatsLoadedStatus = statistics.NewStatsAllEvictedStatus()
+			}
 			lastAnalyzePos.Copy(&col.LastAnalyzePos)
 			table.Columns[hist.ID] = col
 		}
@@ -299,7 +311,7 @@ func (h *Handle) initStatsHistograms4Chunk(is infoschema.InfoSchema, cache util.
 }
 
 func (h *Handle) initStatsHistogramsLite(is infoschema.InfoSchema, cache util.StatsCache) error {
-	sql := "select /*+ ORDER_INDEX(mysql.stats_histograms,tbl)*/ HIGH_PRIORITY table_id, is_index, hist_id, distinct_count, version, null_count, tot_col_size, stats_ver, correlation, flag, last_analyze_pos from mysql.stats_histograms order by table_id"
+	sql := initStatsHistogramsSQLGen(false)
 	rc, err := util.Exec(h.initStatsCtx, sql)
 	if err != nil {
 		return errors.Trace(err)
@@ -322,7 +334,7 @@ func (h *Handle) initStatsHistogramsLite(is infoschema.InfoSchema, cache util.St
 }
 
 func (h *Handle) initStatsHistograms(is infoschema.InfoSchema, cache util.StatsCache) error {
-	sql := "select  /*+ ORDER_INDEX(mysql.stats_histograms,tbl)*/ HIGH_PRIORITY table_id, is_index, hist_id, distinct_count, version, null_count, cm_sketch, tot_col_size, stats_ver, correlation, flag, last_analyze_pos from mysql.stats_histograms order by table_id"
+	sql := initStatsHistogramsSQLGen(false)
 	rc, err := util.Exec(h.initStatsCtx, sql)
 	if err != nil {
 		return errors.Trace(err)
@@ -359,10 +371,7 @@ func (h *Handle) initStatsHistogramsByPaging(is infoschema.InfoSchema, cache uti
 	}()
 
 	sctx := se.(sessionctx.Context)
-	// Why do we need to add `is_index=1` in the SQL?
-	// because it is aligned to the `initStatsTopN` function, which only loads the topn of the index too.
-	// the other will be loaded by sync load.
-	sql := "select HIGH_PRIORITY table_id, is_index, hist_id, distinct_count, version, null_count, cm_sketch, tot_col_size, stats_ver, correlation, flag, last_analyze_pos from mysql.stats_histograms where table_id >= %? and table_id < %? and is_index=1"
+	sql := initStatsHistogramsSQLGen(true)
 	rc, err := util.Exec(sctx, sql, task.StartTid, task.EndTid)
 	if err != nil {
 		return errors.Trace(err)
@@ -676,7 +685,7 @@ func (h *Handle) initStatsBuckets(cache util.StatsCache, totalMemory uint64) err
 			return errors.Trace(err)
 		}
 	} else {
-		sql := "select /*+ ORDER_INDEX(mysql.stats_buckets,tbl)*/ HIGH_PRIORITY table_id, is_index, hist_id, count, repeats, lower_bound, upper_bound, ndv from mysql.stats_buckets order by table_id, is_index, hist_id, bucket_id"
+		sql := "select /*+ ORDER_INDEX(mysql.stats_buckets,tbl)*/ HIGH_PRIORITY table_id, is_index, hist_id, count, repeats, lower_bound, upper_bound, ndv from mysql.stats_buckets where is_index=1 order by table_id, is_index, hist_id, bucket_id"
 		rc, err := util.Exec(h.initStatsCtx, sql)
 		if err != nil {
 			return errors.Trace(err)
@@ -729,7 +738,7 @@ func (h *Handle) initStatsBucketsByPaging(cache util.StatsCache, task initstats.
 		}
 	}()
 	sctx := se.(sessionctx.Context)
-	sql := "select HIGH_PRIORITY table_id, is_index, hist_id, count, repeats, lower_bound, upper_bound, ndv from mysql.stats_buckets where table_id >= %? and table_id < %? order by table_id, is_index, hist_id, bucket_id"
+	sql := "select HIGH_PRIORITY table_id, is_index, hist_id, count, repeats, lower_bound, upper_bound, ndv from mysql.stats_buckets where is_index = 1 and table_id >= %? and table_id < %? order by table_id, is_index, hist_id, bucket_id"
 	rc, err := util.Exec(sctx, sql, task.StartTid, task.EndTid)
 	if err != nil {
 		return errors.Trace(err)

pkg/statistics/handle/handle_hist.go

@@ -23,7 +23,6 @@ import (
 	"github.com/pingcap/errors"
 	"github.com/pingcap/failpoint"
 	"github.com/pingcap/tidb/pkg/config"
-	"github.com/pingcap/tidb/pkg/infoschema"
 	"github.com/pingcap/tidb/pkg/metrics"
 	"github.com/pingcap/tidb/pkg/parser/model"
 	"github.com/pingcap/tidb/pkg/parser/mysql"
@@ -178,7 +177,7 @@ func (h *Handle) removeHistLoadedColumns(neededItems []model.TableItemID) []mode
 			continue
 		}
 		colHist, ok := tbl.Columns[item.ID]
-		if (ok && colHist.IsStatsInitialized() && !colHist.IsFullLoad()) || !ok {
+		if ok && colHist.IsStatsInitialized() && !colHist.IsFullLoad() {
 			remainedItems = append(remainedItems, item)
 		}
 	}
@@ -352,7 +351,7 @@ func (h *Handle) handleOneItemTask(task *NeededItemTask) (err error) {
 var errGetHistMeta = errors.New("fail to get stats version for this histogram")
 
 // readStatsForOneItem reads hist for one column/index, TODO load data via kv-get asynchronously
-func (h *Handle) readStatsForOneItem(sctx sessionctx.Context, item model.TableItemID, w *statsWrapper) (*statsWrapper, error) {
+func (*Handle) readStatsForOneItem(sctx sessionctx.Context, item model.TableItemID, w *statsWrapper) (*statsWrapper, error) {
 	failpoint.Inject("mockReadStatsForOnePanic", nil)
 	failpoint.Inject("mockReadStatsForOneFail", func(val failpoint.Value) {
 		if val.(bool) {
@@ -374,41 +373,9 @@ func (h *Handle) readStatsForOneItem(sctx sessionctx.Context, item model.TableIt
 			return nil, errors.Trace(err)
 		}
 	} else {
-		if c == nil {
-			is := sctx.GetDomainInfoSchema().(infoschema.InfoSchema)
-			tbl, ok := h.TableInfoByID(is, item.TableID)
-			if !ok {
-				return nil, errors.New("no table")
-			}
-			var colInfo *model.ColumnInfo
-			for _, col := range tbl.Meta().Columns {
-				if col.ID == item.ID {
-					colInfo = col
-					break
-				}
-			}
-			if colInfo == nil {
-				return nil, errors.New("no column")
-			}
-			hg, _, _, _, err = storage.HistMetaFromStorageWithHighPriority(sctx, &item, colInfo)
-			if err != nil {
-				return nil, err
-			}
-			if hg != nil {
-				hg, err = storage.HistogramFromStorage(sctx, item.TableID, item.ID, &colInfo.FieldType, hg.NDV, int(isIndexFlag), hg.LastUpdateVersion, hg.NullCount, hg.TotColSize, hg.Correlation)
-				if err != nil {
-					return nil, errors.Trace(err)
-				}
-			}
-			c = &statistics.Column{
-				Info:     colInfo,
-				IsHandle: tbl.Meta().PKIsHandle && mysql.HasPriKeyFlag(colInfo.GetFlag()),
-			}
-		} else {
-			hg, err = storage.HistogramFromStorage(sctx, item.TableID, item.ID, &c.Info.FieldType, c.Histogram.NDV, int(isIndexFlag), c.LastUpdateVersion, c.NullCount, c.TotColSize, c.Correlation)
-			if err != nil {
-				return nil, errors.Trace(err)
-			}
+		hg, err = storage.HistogramFromStorage(sctx, item.TableID, item.ID, &c.Info.FieldType, c.Histogram.NDV, int(isIndexFlag), c.LastUpdateVersion, c.NullCount, c.TotColSize, c.Correlation)
+		if err != nil {
+			return nil, errors.Trace(err)
 		}
 	}
 	var cms *statistics.CMSketch

pkg/statistics/handle/handle_hist_test.go

@@ -403,3 +403,55 @@ func TestSendLoadRequestsWaitTooLong(t *testing.T) {
 		require.Error(t, rs1.Err)
 	}
 }
+
+func TestSyncLoadOnObjectWhichCanNotFoundInStorage(t *testing.T) {
+	store, dom := testkit.CreateMockStoreAndDomain(t)
+	tk := testkit.NewTestKit(t, store)
+	tk.MustExec("use test")
+	tk.MustExec("create table t(a int, b int, c int, primary key(a))")
+	h := dom.StatsHandle()
+	// Skip create table event.
+	<-h.DDLEventCh()
+	tk.MustExec("insert into t values (1,1,1),(2,2,2),(3,3,3)")
+	tk.MustExec("analyze table t columns a, b")
+	tbl, err := dom.InfoSchema().TableByName(model.NewCIStr("test"), model.NewCIStr("t"))
+	require.NoError(t, h.InitStatsLite(dom.InfoSchema()))
+	require.NoError(t, err)
+	require.NotNil(t, tbl)
+	tblInfo := tbl.Meta()
+	statsTbl, ok := h.Get(tblInfo.ID)
+	require.True(t, ok)
+	require.Equal(t, 2, len(statsTbl.Columns))
+	// Do some DDL, one successfully handled by handleDDLEvent, the other not.
+	tk.MustExec("alter table t add column d int default 2")
+	require.NoError(t, h.HandleDDLEvent(<-h.DDLEventCh()))
+	require.NoError(t, h.Update(dom.InfoSchema()))
+	tbl, err = dom.InfoSchema().TableByName(model.NewCIStr("test"), model.NewCIStr("t"))
+	require.NoError(t, err)
+	require.NotNil(t, tbl)
+	tblInfo = tbl.Meta()
+	statsTbl, ok = h.Get(tblInfo.ID)
+	require.True(t, ok)
+	require.Equal(t, 3, len(statsTbl.Columns))
+
+	// Try sync load.
+	tk.MustExec("select * from t where a >= 1 and b = 2 and c = 3 and d = 4")
+	statsTbl, ok = h.Get(tblInfo.ID)
+	require.True(t, ok)
+	require.True(t, statsTbl.Columns[tblInfo.Columns[0].ID].IsFullLoad())
+	require.True(t, statsTbl.Columns[tblInfo.Columns[1].ID].IsFullLoad())
+	require.True(t, statsTbl.Columns[tblInfo.Columns[3].ID].IsFullLoad())
+	require.Nil(t, statsTbl.Columns[tblInfo.Columns[2].ID])
+
+	// Analyze c then test sync load again
+	tk.MustExec("analyze table t columns a, b, c")
+	require.NoError(t, h.InitStatsLite(dom.InfoSchema()))
+	tk.MustExec("select * from t where a >= 1 and b = 2 and c = 3 and d = 4")
+	statsTbl, ok = h.Get(tblInfo.ID)
+	require.True(t, ok)
+	// a, b, d's status is not changed.
+	require.True(t, statsTbl.Columns[tblInfo.Columns[0].ID].IsFullLoad())
+	require.True(t, statsTbl.Columns[tblInfo.Columns[1].ID].IsFullLoad())
+	require.True(t, statsTbl.Columns[tblInfo.Columns[3].ID].IsFullLoad())
+	require.True(t, statsTbl.Columns[tblInfo.Columns[2].ID].IsFullLoad())
+}

pkg/statistics/handle/handletest/handle_test.go

@@ -1606,7 +1606,12 @@ func TestInitStatsLite(t *testing.T) {
 	statsTbl1 := h.GetTableStats(tblInfo)
 	checkAllEvicted(t, statsTbl1)
 	internal.AssertTableEqual(t, statsTbl0, statsTbl1)
-
+	for _, col := range statsTbl1.Columns {
+		require.Equal(t, int64(statistics.Version2), col.StatsVer)
+	}
+	for _, idx := range statsTbl1.Indices {
+		require.Equal(t, int64(statistics.Version2), idx.StatsVer)
+	}
 	// async stats load
 	tk.MustExec("set @@tidb_stats_load_sync_wait = 0")
 	tk.MustExec("explain select * from t where b > 1")

pkg/statistics/handle/handletest/initstats/load_stats_test.go

@@ -86,7 +86,7 @@ func testConcurrentlyInitStats(t *testing.T) {
 		tk.MustQuery(fmt.Sprintf("explain select * from t%v where b = 1", i)).CheckNotContain("pseudo")
 	}
 	for i := 1; i < 10; i++ {
-		tk.MustQuery(fmt.Sprintf("explain select * from t%v where c = 1", i)).CheckNotContain("pseudo")
+		tk.MustQuery(fmt.Sprintf("explain select * from t%v where c >= 1", i)).CheckNotContain("pseudo")
 	}
 	for i := 1; i < 10; i++ {
 		tbl, err := is.TableByName(model.NewCIStr("test"), model.NewCIStr(fmt.Sprintf("t%v", i)))

pkg/statistics/handle/handletest/statstest/stats_test.go

@@ -270,6 +270,9 @@ func TestInitStats(t *testing.T) {
 	require.NoError(t, h.Update(is))
 	// Index and pk are loaded.
 	needed := fmt.Sprintf(`Table:%v RealtimeCount:6
+column:1 ndv:6 totColSize:0
+column:2 ndv:6 totColSize:6
+column:3 ndv:6 totColSize:6
 index:1 ndv:6
 num: 1 lower_bound: 1 upper_bound: 1 repeats: 1 ndv: 0
 num: 1 lower_bound: 2 upper_bound: 2 repeats: 1 ndv: 0
@@ -312,6 +315,12 @@ func TestInitStats2(t *testing.T) {
 	h.Clear()
 	require.NoError(t, h.Update(is))
 	table1 := h.GetTableStats(tbl.Meta())
+	// stats of pk will be loaded.
+	require.Equal(t, true, table0.Columns[1].IsAllEvicted())
+	require.Equal(t, true, table1.Columns[1].IsFullLoad())
+	delete(table0.Columns, 1)
+	delete(table1.Columns, 1)
+	// result part is not changed.
 	internal.AssertTableEqual(t, table0, table1)
 	h.SetLease(0)
 }