Merge #76866

76866: opt: add use-multi-col-stats test flag r=mgartner a=mgartner

A new optimizer test flag, `use-multi-col-stats`, has been added to allow
disabling multi-column stats for a given test. This can be helpful when
debugging issues with multi-column stats.

Release note: None

Co-authored-by: Marcus Gartner <[email protected]>

pkg/sql/opt/testutils/opttester/opt_tester.go

@@ -245,14 +245,18 @@ type Flags struct {
 
 	// QueryArgs are values for placeholders, used for assign-placeholders-*.
 	QueryArgs []string
+
+	// UseMultiColStats is the value for SessionData.OptimizerUseMultiColStats.
+	// It defaults to true in New.
+	UseMultiColStats bool
 }
 
 // New constructs a new instance of the OptTester for the given SQL statement.
 // Metadata used by the SQL query is accessed via the catalog.
 func New(catalog cat.Catalog, sql string) *OptTester {
 	ctx := context.Background()
 	ot := &OptTester{
-		Flags:   Flags{JoinLimit: opt.DefaultJoinOrderLimit},
+		Flags:   Flags{JoinLimit: opt.DefaultJoinOrderLimit, UseMultiColStats: true},
 		catalog: catalog,
 		sql:     sql,
 		ctx:     ctx,
@@ -269,7 +273,6 @@ func New(catalog cat.Catalog, sql string) *OptTester {
 	ot.evalCtx.SessionData().Database = "defaultdb"
 	ot.evalCtx.SessionData().ZigzagJoinEnabled = true
 	ot.evalCtx.SessionData().OptimizerUseHistograms = true
-	ot.evalCtx.SessionData().OptimizerUseMultiColStats = true
 	ot.evalCtx.SessionData().LocalityOptimizedSearch = true
 	ot.evalCtx.SessionData().ReorderJoinsLimit = opt.DefaultJoinOrderLimit
 	ot.evalCtx.SessionData().InsertFastPath = true
@@ -517,6 +520,7 @@ func (ot *OptTester) RunCommand(tb testing.TB, d *datadriven.TestData) string {
 	ot.evalCtx.SessionData().PreferLookupJoinsForFKs = ot.Flags.PreferLookupJoinsForFKs
 	ot.evalCtx.SessionData().PropagateInputOrdering = ot.Flags.PropagateInputOrdering
 	ot.evalCtx.SessionData().NullOrderedLast = ot.Flags.NullOrderedLast
+	ot.evalCtx.SessionData().OptimizerUseMultiColStats = ot.Flags.UseMultiColStats
 
 	ot.evalCtx.TestingKnobs.OptimizerCostPerturbation = ot.Flags.PerturbCost
 	ot.evalCtx.Locality = ot.Flags.Locality
@@ -1064,6 +1068,16 @@ func (f *Flags) Set(arg datadriven.CmdArg) error {
 	case "propagate-input-ordering":
 		f.PropagateInputOrdering = true
 
+	case "use-multi-col-stats":
+		if len(arg.Vals) != 1 {
+			return fmt.Errorf("use-multi-col-stats requires a single argument")
+		}
+		b, err := strconv.ParseBool(arg.Vals[0])
+		if err != nil {
+			return errors.Wrap(err, "use-multi-col-stats")
+		}
+		f.UseMultiColStats = b
+
 	default:
 		return fmt.Errorf("unknown argument: %s", arg.Key)
 	}

pkg/sql/opt/testutils/opttester/testdata/use-multi-col-stats

@@ -0,0 +1,44 @@
+exec-ddl
+CREATE TABLE rides (
+    id UUID NOT NULL,
+    city VARCHAR NOT NULL,
+    vehicle_city VARCHAR NULL,
+    rider_id UUID NULL,
+    vehicle_id UUID NULL,
+    start_address VARCHAR NULL,
+    end_address VARCHAR NULL,
+    start_time TIMESTAMP NULL,
+    end_time TIMESTAMP NULL,
+    revenue DECIMAL(10,2) NULL,
+    CONSTRAINT "primary" PRIMARY KEY (city ASC, id ASC),
+    INDEX rides_auto_index_fk_city_ref_users (city ASC, rider_id ASC),
+    INDEX rides_auto_index_fk_vehicle_city_ref_vehicles (vehicle_city ASC, vehicle_id ASC),
+    FAMILY "primary" (id, city, vehicle_city, rider_id, vehicle_id, start_address, end_address, start_time, end_time, revenue),
+    CONSTRAINT check_vehicle_city_city CHECK (vehicle_city = city)
+)
+----
+
+inject-stats file=inject-stats.json table=rides
+----
+
+opt format=show-stats use-multi-col-stats=true
+SELECT * FROM rides WHERE city = 'rome' and id = '17198184-b24f-4aa8-9933-64a72ff6665f'
+----
+scan rides
+ ├── columns: id:1(uuid!null) city:2(varchar!null) vehicle_city:3(varchar) rider_id:4(uuid) vehicle_id:5(uuid) start_address:6(varchar) end_address:7(varchar) start_time:8(timestamp) end_time:9(timestamp) revenue:10(decimal)
+ ├── constraint: /2/1: [/'rome'/'17198184-b24f-4aa8-9933-64a72ff6665f' - /'rome'/'17198184-b24f-4aa8-9933-64a72ff6665f']
+ ├── cardinality: [0 - 1]
+ ├── stats: [rows=0.9111111, distinct(1)=0.911111, null(1)=0, avgsize(1)=1, distinct(2)=0.911111, null(2)=0, avgsize(2)=6, distinct(1,2)=0.911111, null(1,2)=0, avgsize(1,2)=8]
+ ├── key: ()
+ └── fd: ()-->(1-10)
+
+opt format=show-stats use-multi-col-stats=false
+SELECT * FROM rides WHERE city = 'rome' and id = '17198184-b24f-4aa8-9933-64a72ff6665f'
+----
+scan rides
+ ├── columns: id:1(uuid!null) city:2(varchar!null) vehicle_city:3(varchar) rider_id:4(uuid) vehicle_id:5(uuid) start_address:6(varchar) end_address:7(varchar) start_time:8(timestamp) end_time:9(timestamp) revenue:10(decimal)
+ ├── constraint: /2/1: [/'rome'/'17198184-b24f-4aa8-9933-64a72ff6665f' - /'rome'/'17198184-b24f-4aa8-9933-64a72ff6665f']
+ ├── cardinality: [0 - 1]
+ ├── stats: [rows=0.1111111, distinct(1)=0.111111, null(1)=0, avgsize(1)=1, distinct(2)=0.111111, null(2)=0, avgsize(2)=6]
+ ├── key: ()
+ └── fd: ()-->(1-10)