planner: respect tidb_analyze_column_options when build analyze plan

pkg/executor/test/analyzetest/analyze_test.go

@@ -1040,7 +1040,6 @@ func TestSavedAnalyzeColumnOptions(t *testing.T) {
 	defer func() {
 		tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal4))
 	}()
-	tk.MustExec("set global tidb_enable_column_tracking = 1")
 
 	tk.MustExec("use test")
 	tk.MustExec("set @@session.tidb_analyze_version = 2")
@@ -1151,7 +1150,6 @@ func TestAnalyzeColumnsWithPrimaryKey(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where a > 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1219,7 +1217,6 @@ func TestAnalyzeColumnsWithIndex(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where c > 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1296,7 +1293,6 @@ func TestAnalyzeColumnsWithClusteredIndex(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where c > 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1378,7 +1374,6 @@ func TestAnalyzeColumnsWithDynamicPartitionTable(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where a < 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1502,7 +1497,6 @@ func TestAnalyzeColumnsWithStaticPartitionTable(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where a < 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1606,7 +1600,6 @@ func TestAnalyzeColumnsWithExtendedStats(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where b > 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1676,7 +1669,6 @@ func TestAnalyzeColumnsWithVirtualColumnIndex(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where b > 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1759,7 +1751,6 @@ func TestAnalyzeColumnsAfterAnalyzeAll(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where b > 1")
 				require.NoError(t, h.DumpColStatsUsageToKV())
 				rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null").Rows()
@@ -1835,10 +1826,8 @@ func TestAnalyzeColumnsErrorAndWarning(t *testing.T) {
 	tk.MustExec("analyze table t predicate columns")
 	tk.MustQuery("show warnings").Sort().Check(testkit.Rows(
 		`Note 1105 Analyze use auto adjusted sample rate 1.000000 for table test.t, reason to use this rate is "use min(1, 110000/10000) as the sample-rate=1"`,
-		"Warning 1105 No predicate column has been collected yet for table test.t so all columns are analyzed",
+		"Warning 1105 No predicate column has been collected yet for table test.t, so only indexes and the columns composing the indexes will be analyzed",
 	))
-	rows := tk.MustQuery("show column_stats_usage where db_name = 'test' and table_name = 't' and last_analyzed_at is not null").Rows()
-	require.Equal(t, 2, len(rows))
 
 	for _, val := range []model.ColumnChoice{model.ColumnList, model.PredicateColumns} {
 		func(choice model.ColumnChoice) {
@@ -1853,7 +1842,6 @@ func TestAnalyzeColumnsErrorAndWarning(t *testing.T) {
 				defer func() {
 					tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 				}()
-				tk.MustExec("set global tidb_enable_column_tracking = 1")
 				tk.MustExec("select * from t where b > 1")
 				require.NoError(t, dom.StatsHandle().DumpColStatsUsageToKV())
 				tk.MustExec("analyze table t predicate columns")
@@ -2167,7 +2155,6 @@ func TestShowAanalyzeStatusJobInfo(t *testing.T) {
 	}
 	checkJobInfo("analyze table columns b, c, d with 2 buckets, 2 topn, 1 samplerate")
 	tk.MustExec("set global tidb_persist_analyze_options = 1")
-	tk.MustExec("set global tidb_enable_column_tracking = 1")
 	tk.MustExec("select * from t where c > 1")
 	h := dom.StatsHandle()
 	require.NoError(t, h.DumpColStatsUsageToKV())

pkg/planner/core/planbuilder.go

@@ -1962,6 +1962,7 @@ func (b *PlanBuilder) getMustAnalyzedColumns(tbl *ast.TableName, cols *calcOnceM
 	return cols.data, nil
 }
 
+// getPredicateColumns gets the columns used in predicates.
 func (b *PlanBuilder) getPredicateColumns(tbl *ast.TableName, cols *calcOnceMap) (map[int64]struct{}, error) {
 	// Already calculated in the previous call.
 	if cols.calculated {
@@ -1976,10 +1977,13 @@ func (b *PlanBuilder) getPredicateColumns(tbl *ast.TableName, cols *calcOnceMap)
 		return nil, err
 	}
 	if len(colList) == 0 {
-		b.ctx.GetSessionVars().StmtCtx.AppendWarning(errors.NewNoStackErrorf("No predicate column has been collected yet for table %s.%s so all columns are analyzed", tbl.Schema.L, tbl.Name.L))
-		for _, colInfo := range tblInfo.Columns {
-			cols.data[colInfo.ID] = struct{}{}
-		}
+		b.ctx.GetSessionVars().StmtCtx.AppendWarning(
+			errors.NewNoStackErrorf(
+				"No predicate column has been collected yet for table %s.%s, so only indexes and the columns composing the indexes will be analyzed",
+				tbl.Schema.L,
+				tbl.Name.L,
+			),
+		)
 	} else {
 		for _, id := range colList {
 			cols.data[id] = struct{}{}
@@ -2017,22 +2021,40 @@ func (b *PlanBuilder) getFullAnalyzeColumnsInfo(
 	}
 
 	switch columnChoice {
-	case model.DefaultChoice, model.AllColumns:
-		return tbl.TableInfo.Columns, nil, nil
-	case model.PredicateColumns:
-		if mustAllColumns {
+	case model.DefaultChoice:
+		columnOptions := variable.AnalyzeColumnOptions.Load()
+		switch columnOptions {
+		case model.AllColumns.String():
+			return tbl.TableInfo.Columns, nil, nil
+		case model.PredicateColumns.String():
+			columns, err := b.getColumnsBasedOnPredicateColumns(
+				tbl,
+				predicateCols,
+				mustAnalyzedCols,
+				mustAllColumns,
+			)
+			if err != nil {
+				return nil, nil, err
+			}
+			return columns, nil, nil
+		default:
+			// Usually, this won't happen.
+			logutil.BgLogger().Warn("Unknown default column choice, analyze all columns", zap.String("choice", columnOptions))
 			return tbl.TableInfo.Columns, nil, nil
 		}
-		predicate, err := b.getPredicateColumns(tbl, predicateCols)
-		if err != nil {
-			return nil, nil, err
-		}
-		mustAnalyzed, err := b.getMustAnalyzedColumns(tbl, mustAnalyzedCols)
+	case model.AllColumns:
+		return tbl.TableInfo.Columns, nil, nil
+	case model.PredicateColumns:
+		columns, err := b.getColumnsBasedOnPredicateColumns(
+			tbl,
+			predicateCols,
+			mustAnalyzedCols,
+			mustAllColumns,
+		)
 		if err != nil {
 			return nil, nil, err
 		}
-		colSet := combineColumnSets(predicate, mustAnalyzed)
-		return getColumnListFromSet(tbl.TableInfo.Columns, colSet), nil, nil
+		return columns, nil, nil
 	case model.ColumnList:
 		colSet := getColumnSetFromSpecifiedCols(specifiedCols)
 		mustAnalyzed, err := b.getMustAnalyzedColumns(tbl, mustAnalyzedCols)
@@ -2058,6 +2080,26 @@ func (b *PlanBuilder) getFullAnalyzeColumnsInfo(
 	return nil, nil, nil
 }
 
+func (b *PlanBuilder) getColumnsBasedOnPredicateColumns(
+	tbl *ast.TableName,
+	predicateCols, mustAnalyzedCols *calcOnceMap,
+	rewriteAllStatsNeeded bool,
+) ([]*model.ColumnInfo, error) {
+	if rewriteAllStatsNeeded {
+		return tbl.TableInfo.Columns, nil
+	}
+	predicate, err := b.getPredicateColumns(tbl, predicateCols)
+	if err != nil {
+		return nil, err
+	}
+	mustAnalyzed, err := b.getMustAnalyzedColumns(tbl, mustAnalyzedCols)
+	if err != nil {
+		return nil, err
+	}
+	colSet := combineColumnSets(predicate, mustAnalyzed)
+	return getColumnListFromSet(tbl.TableInfo.Columns, colSet), nil
+}
+
 // Helper function to combine two column sets.
 func combineColumnSets(sets ...map[int64]struct{}) map[int64]struct{} {
 	result := make(map[int64]struct{})

pkg/statistics/handle/autoanalyze/BUILD.bazel

@@ -35,7 +35,7 @@ go_test(
     timeout = "short",
     srcs = ["autoanalyze_test.go"],
     flaky = True,
-    shard_count = 12,
+    shard_count = 13,
     deps = [
         ":autoanalyze",
         "//pkg/domain/infosync",

pkg/statistics/handle/autoanalyze/autoanalyze_test.go

@@ -90,6 +90,45 @@ func TestAutoAnalyzeLockedTable(t *testing.T) {
 	require.True(t, dom.StatsHandle().HandleAutoAnalyze())
 }
 
+func TestAutoAnalyzeWithPredicateColumns(t *testing.T) {
+	// Create a table and add it to stats cache.
+	store, dom := testkit.CreateMockStoreAndDomain(t)
+	tk := testkit.NewTestKit(t, store)
+	tk.MustExec("use test")
+	tk.MustExec("create table t (a int, b int)")
+	tk.MustExec("insert into t values (1, 1)")
+	tk.MustQuery("select * from t where a > 0").Check(testkit.Rows("1 1"))
+	h := dom.StatsHandle()
+	err := h.HandleDDLEvent(<-h.DDLEventCh())
+	require.NoError(t, err)
+	require.NoError(t, h.DumpColStatsUsageToKV())
+	require.NoError(t, h.DumpStatsDeltaToKV(true))
+	is := dom.InfoSchema()
+	require.NoError(t, h.Update(is))
+	exec.AutoAnalyzeMinCnt = 0
+	defer func() {
+		exec.AutoAnalyzeMinCnt = 1000
+	}()
+
+	// Check column_stats_usage.
+	rows := tk.MustQuery(
+		"show column_stats_usage where db_name = 'test' and table_name = 't' and last_used_at is not null",
+	).Rows()
+	require.Equal(t, 1, len(rows))
+	require.Equal(t, "a", rows[0][3])
+
+	// Set tidb_analyze_column_options to PREDICATE.
+	tk.MustExec("set global tidb_analyze_column_options='PREDICATE'")
+
+	// Trigger auto analyze.
+	require.True(t, dom.StatsHandle().HandleAutoAnalyze())
+
+	// Check analyze jobs.
+	tk.MustQuery("select table_name, job_info from mysql.analyze_jobs order by id desc limit 1").Check(
+		testkit.Rows("t auto analyze table columns a with 256 buckets, 100 topn, 1 samplerate"),
+	)
+}
+
 func TestDisableAutoAnalyze(t *testing.T) {
 	store, dom := testkit.CreateMockStoreAndDomain(t)
 	tk := testkit.NewTestKit(t, store)

pkg/statistics/handle/updatetest/update_test.go

@@ -905,7 +905,6 @@ func TestDumpColumnStatsUsage(t *testing.T) {
 	defer func() {
 		tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal))
 	}()
-	tk.MustExec("set global tidb_enable_column_tracking = 1")
 
 	h := dom.StatsHandle()
 	tk.MustExec("use test")
@@ -987,7 +986,6 @@ func TestCollectPredicateColumnsFromExecute(t *testing.T) {
 			defer func() {
 				tk.MustExec(fmt.Sprintf("set global tidb_enable_column_tracking = %v", originalVal2))
 			}()
-			tk.MustExec("set global tidb_enable_column_tracking = 1")
 
 			h := dom.StatsHandle()
 			tk.MustExec("use test")

pkg/statistics/handle/usage/BUILD.bazel

@@ -40,7 +40,7 @@ go_test(
     ],
     embed = [":usage"],
     flaky = True,
-    shard_count = 3,
+    shard_count = 10,
     deps = [
         "//pkg/infoschema",
         "//pkg/parser/model",