infoschema: optimize 'select count(*) from information_schema.tables' for v2 (#55574)

close #55515

Author: tiancaiamao

pkg/executor/infoschema_reader.go

@@ -689,15 +689,25 @@ func (e *memtableRetriever) setDataFromOneTable(
 	return rows, nil
 }
 
-func (e *memtableRetriever) setDataFromTables(ctx context.Context, sctx sessionctx.Context) error {
-	useStatsCache := e.updateStatsCacheIfNeed()
-	checker := privilege.GetPrivilegeManager(sctx)
+func onlySchemaOrTableColumns(columns []*model.ColumnInfo) bool {
+	if len(columns) <= 3 {
+		for _, colInfo := range columns {
+			switch colInfo.Name.L {
+			case "table_schema":
+			case "table_name":
+			case "table_catalog":
+			default:
+				return false
+			}
+		}
+		return true
+	}
+	return false
+}
 
+func (e *memtableRetriever) setDataFromTables(ctx context.Context, sctx sessionctx.Context) error {
 	var rows [][]types.Datum
-	loc := sctx.GetSessionVars().TimeZone
-	if loc == nil {
-		loc = time.Local
-	}
+	checker := privilege.GetPrivilegeManager(sctx)
 	ex, ok := e.extractor.(*plannercore.InfoSchemaTablesExtractor)
 	if !ok {
 		return errors.Errorf("wrong extractor type: %T, expected InfoSchemaTablesExtractor", e.extractor)
@@ -706,10 +716,82 @@ func (e *memtableRetriever) setDataFromTables(ctx context.Context, sctx sessionc
 		return nil
 	}
 
+	// Special optimize for queries on infoschema v2 like:
+	//     select count(table_schema) from INFORMATION_SCHEMA.TABLES
+	//     select count(*) from INFORMATION_SCHEMA.TABLES
+	//     select table_schema, table_name from INFORMATION_SCHEMA.TABLES
+	// column pruning in general is not supported here.
+	if onlySchemaOrTableColumns(e.columns) {
+		is := e.is
+		if raw, ok := is.(*infoschema.SessionExtendedInfoSchema); ok {
+			is = raw.InfoSchema
+		}
+		v2, ok := is.(interface {
+			IterateAllTableItems(visit func(infoschema.TableItem) bool)
+		})
+		if ok {
+			if x := ctx.Value("cover-check"); x != nil {
+				// The interface assertion is too tricky, so we add test to cover here.
+				// To ensure that if implementation changes one day, we can catch it.
+				slot := x.(*bool)
+				*slot = true
+			}
+			v2.IterateAllTableItems(func(t infoschema.TableItem) bool {
+				if !ex.HasTableName(t.TableName.L) {
+					return true
+				}
+				if !ex.HasTableSchema(t.DBName.L) {
+					return true
+				}
+				if checker != nil && !checker.RequestVerification(sctx.GetSessionVars().ActiveRoles, t.DBName.L, t.TableName.L, "", mysql.SelectPriv) {
+					return true
+				}
+
+				record := types.MakeDatums(
+					infoschema.CatalogVal, // TABLE_CATALOG
+					t.DBName.O,            // TABLE_SCHEMA
+					t.TableName.O,         // TABLE_NAME
+					nil,                   // TABLE_TYPE
+					nil,                   // ENGINE
+					nil,                   // VERSION
+					nil,                   // ROW_FORMAT
+					nil,                   // TABLE_ROWS
+					nil,                   // AVG_ROW_LENGTH
+					nil,                   // DATA_LENGTH
+					nil,                   // MAX_DATA_LENGTH
+					nil,                   // INDEX_LENGTH
+					nil,                   // DATA_FREE
+					nil,                   // AUTO_INCREMENT
+					nil,                   // CREATE_TIME
+					nil,                   // UPDATE_TIME
+					nil,                   // CHECK_TIME
+					nil,                   // TABLE_COLLATION
+					nil,                   // CHECKSUM
+					nil,                   // CREATE_OPTIONS
+					nil,                   // TABLE_COMMENT
+					nil,                   // TIDB_TABLE_ID
+					nil,                   // TIDB_ROW_ID_SHARDING_INFO
+					nil,                   // TIDB_PK_TYPE
+					nil,                   // TIDB_PLACEMENT_POLICY_NAME
+				)
+				rows = append(rows, record)
+				return true
+			})
+			e.rows = rows
+			return nil
+		}
+	}
+
+	// Normal code path.
 	schemas, tables, err := ex.ListSchemasAndTables(ctx, e.is)
 	if err != nil {
 		return errors.Trace(err)
 	}
+	useStatsCache := e.updateStatsCacheIfNeed()
+	loc := sctx.GetSessionVars().TimeZone
+	if loc == nil {
+		loc = time.Local
+	}
 	for i, table := range tables {
 		rows, err = e.setDataFromOneTable(sctx, loc, checker, schemas[i], table, rows, useStatsCache)
 		if err != nil {

pkg/executor/infoschema_reader_test.go

@@ -957,3 +957,32 @@ func TestInfoSchemaConditionWorks(t *testing.T) {
 	rows = tk.MustQuery("select * from information_schema.partitions where table_schema = 'db_no_partition' and (partition_name is NULL or partition_name = 'p0');").Rows()
 	require.Equal(t, 2, len(rows))
 }
+
+func TestInfoschemaTablesSpecialOptimizationCovered(t *testing.T) {
+	store := testkit.CreateMockStore(t)
+	tk := testkit.NewTestKit(t, store)
+
+	for _, testCase := range []struct {
+		sql    string
+		expect bool
+	}{
+		{"select table_name, table_schema from information_schema.tables", true},
+		{"select table_name from information_schema.tables", true},
+		{"select table_name from information_schema.tables where table_schema = 'test'", true},
+		{"select table_schema from information_schema.tables", true},
+		{"select count(table_schema) from information_schema.tables", true},
+		{"select count(table_name) from information_schema.tables", true},
+		{"select count(table_rows) from information_schema.tables", false},
+		{"select count(1) from information_schema.tables", true},
+		{"select count(*) from information_schema.tables", true},
+		{"select count(1) from (select table_name from information_schema.tables) t", true},
+		{"select * from information_schema.tables", false},
+		{"select table_name, table_catalog from information_schema.tables", true},
+		{"select table_name, table_rows from information_schema.tables", false},
+	} {
+		var covered bool
+		ctx := context.WithValue(context.Background(), "cover-check", &covered)
+		tk.MustQueryWithContext(ctx, testCase.sql)
+		require.Equal(t, testCase.expect, covered, testCase.sql)
+	}
+}

pkg/infoschema/infoschema_v2.go

@@ -645,6 +645,34 @@ func (is *infoschemaV2) TableByID(ctx context.Context, id int64) (val table.Tabl
 	return ret, true
 }
 
+// TableItem is exported from tableItem.
+type TableItem struct {
+	DBName    model.CIStr
+	TableName model.CIStr
+}
+
+// IterateAllTableItems is used for special performance optimization.
+// Used by executor/infoschema_reader.go to handle reading from INFORMATION_SCHEMA.TABLES.
+func (is *infoschemaV2) IterateAllTableItems(visit func(TableItem) bool) {
+	pivot, ok := is.byName.Max()
+	if !ok {
+		return
+	}
+	if !visit(TableItem{DBName: pivot.dbName, TableName: pivot.tableName}) {
+		return
+	}
+	is.byName.Descend(pivot, func(item tableItem) bool {
+		if pivot.dbName == item.dbName && pivot.tableName == item.tableName {
+			return true // skip MVCC version
+		}
+		pivot = item
+		if !item.tomb {
+			return visit(TableItem{DBName: item.dbName, TableName: item.tableName})
+		}
+		return true
+	})
+}
+
 // IsSpecialDB tells whether the database is a special database.
 func IsSpecialDB(dbName string) bool {
 	return dbName == util.InformationSchemaName.L ||

pkg/planner/core/memtable_infoschema_extractor.go

@@ -260,6 +260,16 @@ func NewInfoSchemaTablesExtractor() *InfoSchemaTablesExtractor {
 	return e
 }
 
+// HasTableName returns true if table name is specified in predicates.
+func (e *InfoSchemaTablesExtractor) HasTableName(name string) bool {
+	return !e.filter(TableName, name)
+}
+
+// HasTableSchema returns true if table schema is specified in predicates.
+func (e *InfoSchemaTablesExtractor) HasTableSchema(name string) bool {
+	return !e.filter(TableSchema, name)
+}
+
 // InfoSchemaViewsExtractor is the predicate extractor for information_schema.views.
 type InfoSchemaViewsExtractor struct {
 	InfoSchemaBaseExtractor

tests/integrationtest/r/executor/infoschema_reader.result

@@ -420,5 +420,22 @@ select * from information_schema.table_constraints where table_schema = 'executo
 CONSTRAINT_CATALOG	CONSTRAINT_SCHEMA	CONSTRAINT_NAME	TABLE_SCHEMA	TABLE_NAME	CONSTRAINT_TYPE
 select * from information_schema.table_constraints where table_schema = 'executor__infoschema_reader' and CONSTRAINT_NAME = 'c1';
 CONSTRAINT_CATALOG	CONSTRAINT_SCHEMA	CONSTRAINT_NAME	TABLE_SCHEMA	TABLE_NAME	CONSTRAINT_TYPE
-def	executor__infoschema_reader	PRIMARY	executor__infoschema_reader	t	PRIMARY KEY
-def	executor__infoschema_reader	PRIMARY	executor__infoschema_reader	t_int	PRIMARY KEY
+select TABLE_CATALOG,TABLE_SCHEMA,TABLE_NAME,TABLE_TYPE,ENGINE,VERSION from information_schema.tables where table_name = 't';
+TABLE_CATALOG	TABLE_SCHEMA	TABLE_NAME	TABLE_TYPE	ENGINE	VERSION
+def	executor__infoschema_reader	t	BASE TABLE	InnoDB	10
+select table_name, table_schema from information_schema.tables where table_name = 't';
+table_name	table_schema
+t	executor__infoschema_reader
+select table_name from information_schema.tables where table_name = 't';
+table_name
+t
+explain format='brief' select table_name, table_schema from information_schema.tables;
+id	estRows	task	access object	operator info
+Projection	10000.00	root		Column#3, Column#2
+└─MemTableScan	10000.00	root	table:TABLES	
+select count(*) from information_schema.tables where table_name = 't';
+count(*)
+1
+select count(table_name) from information_schema.tables where table_name = 't';
+count(table_name)
+1

tests/integrationtest/t/executor/infoschema_reader.test

@@ -307,3 +307,11 @@ select * from information_schema.table_constraints where table_schema = 'executo
 select * from information_schema.table_constraints where table_schema = 'executor__infoschema_reader' and table_name = 'non_exist';
 --sorted_result
 select * from information_schema.table_constraints where table_schema = 'executor__infoschema_reader' and CONSTRAINT_NAME = 'c1';
+
+# TestTables
+select TABLE_CATALOG,TABLE_SCHEMA,TABLE_NAME,TABLE_TYPE,ENGINE,VERSION from information_schema.tables where table_name = 't';
+select table_name, table_schema from information_schema.tables where table_name = 't';
+select table_name from information_schema.tables where table_name = 't';
+explain format='brief' select table_name, table_schema from information_schema.tables;
+select count(*) from information_schema.tables where table_name = 't';
+select count(table_name) from information_schema.tables where table_name = 't';