planner: add new logical rule for constant propagation (#46544)

close #15082
pingcap · Sep 11, 2023 · 02ceded · 02ceded
1 parent e9ef984
commit 02ceded
Show file tree

Hide file tree

Showing 18 changed files with 887 additions and 92 deletions.
diff --git a/cmd/explaintest/r/explain_cte.result b/cmd/explaintest/r/explain_cte.result
@@ -511,17 +511,17 @@ a	b
 explain insert into t1  select t1.a, t1.b  from  t1 inner join (select t2.c from t2 inner join (with temp as (select e from t3 where t3.f = 1234) select e from temp) tt on t2.d = tt.e) t on t1.a = t.c;
 id	estRows	task	access object	operator info
 Insert_1	N/A	root		N/A
-└─HashJoin_25	15.61	root		inner join, equal:[eq(explain_cte.t2.c, explain_cte.t1.a)]
-  ├─HashJoin_27(Build)	12.49	root		inner join, equal:[eq(explain_cte.t3.e, explain_cte.t2.d)]
-  │ ├─TableReader_30(Build)	9.99	root		data:Selection_29
-  │ │ └─Selection_29	9.99	cop[tikv]		eq(explain_cte.t3.f, 1234), not(isnull(explain_cte.t3.e))
-  │ │   └─TableFullScan_28	10000.00	cop[tikv]	table:t3	keep order:false, stats:pseudo
-  │ └─TableReader_33(Probe)	9980.01	root		data:Selection_32
-  │   └─Selection_32	9980.01	cop[tikv]		not(isnull(explain_cte.t2.c)), not(isnull(explain_cte.t2.d))
-  │     └─TableFullScan_31	10000.00	cop[tikv]	table:t2	keep order:false, stats:pseudo
-  └─TableReader_36(Probe)	9990.00	root		data:Selection_35
-    └─Selection_35	9990.00	cop[tikv]		not(isnull(explain_cte.t1.a))
-      └─TableFullScan_34	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
+└─HashJoin_26	15.61	root		inner join, equal:[eq(explain_cte.t2.c, explain_cte.t1.a)]
+  ├─HashJoin_28(Build)	12.49	root		inner join, equal:[eq(explain_cte.t3.e, explain_cte.t2.d)]
+  │ ├─TableReader_31(Build)	9.99	root		data:Selection_30
+  │ │ └─Selection_30	9.99	cop[tikv]		eq(explain_cte.t3.f, 1234), not(isnull(explain_cte.t3.e))
+  │ │   └─TableFullScan_29	10000.00	cop[tikv]	table:t3	keep order:false, stats:pseudo
+  │ └─TableReader_34(Probe)	9980.01	root		data:Selection_33
+  │   └─Selection_33	9980.01	cop[tikv]		not(isnull(explain_cte.t2.c)), not(isnull(explain_cte.t2.d))
+  │     └─TableFullScan_32	10000.00	cop[tikv]	table:t2	keep order:false, stats:pseudo
+  └─TableReader_37(Probe)	9990.00	root		data:Selection_36
+    └─Selection_36	9990.00	cop[tikv]		not(isnull(explain_cte.t1.a))
+      └─TableFullScan_35	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
 insert into t1  select t1.a, t1.b  from  t1 inner join (select t2.c from t2 inner join (with temp as (select e from t3 where t3.f = 1234) select e from temp) tt on t2.d = tt.e) t on t1.a = t.c;
 select * from t1;
 a	b
@@ -530,18 +530,18 @@ a	b
 explain delete from  t1 using t1 inner join (select t2.c from t2 inner join (with temp as (select e from t3 where t3.f = 1234) select e from temp) tt on t2.d = tt.e) t on t1.a = t.c;
 id	estRows	task	access object	operator info
 Delete_17	N/A	root		N/A
-└─Projection_22	15.61	root		explain_cte.t1.a, explain_cte.t1.b, explain_cte.t1._tidb_rowid, explain_cte.t2.c
-  └─HashJoin_24	15.61	root		inner join, equal:[eq(explain_cte.t2.c, explain_cte.t1.a)]
-    ├─HashJoin_26(Build)	12.49	root		inner join, equal:[eq(explain_cte.t3.e, explain_cte.t2.d)]
-    │ ├─TableReader_29(Build)	9.99	root		data:Selection_28
-    │ │ └─Selection_28	9.99	cop[tikv]		eq(explain_cte.t3.f, 1234), not(isnull(explain_cte.t3.e))
-    │ │   └─TableFullScan_27	10000.00	cop[tikv]	table:t3	keep order:false, stats:pseudo
-    │ └─TableReader_32(Probe)	9980.01	root		data:Selection_31
-    │   └─Selection_31	9980.01	cop[tikv]		not(isnull(explain_cte.t2.c)), not(isnull(explain_cte.t2.d))
-    │     └─TableFullScan_30	10000.00	cop[tikv]	table:t2	keep order:false, stats:pseudo
-    └─TableReader_35(Probe)	9990.00	root		data:Selection_34
-      └─Selection_34	9990.00	cop[tikv]		not(isnull(explain_cte.t1.a))
-        └─TableFullScan_33	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
+└─Projection_23	15.61	root		explain_cte.t1.a, explain_cte.t1.b, explain_cte.t1._tidb_rowid, explain_cte.t2.c
+  └─HashJoin_25	15.61	root		inner join, equal:[eq(explain_cte.t2.c, explain_cte.t1.a)]
+    ├─HashJoin_27(Build)	12.49	root		inner join, equal:[eq(explain_cte.t3.e, explain_cte.t2.d)]
+    │ ├─TableReader_30(Build)	9.99	root		data:Selection_29
+    │ │ └─Selection_29	9.99	cop[tikv]		eq(explain_cte.t3.f, 1234), not(isnull(explain_cte.t3.e))
+    │ │   └─TableFullScan_28	10000.00	cop[tikv]	table:t3	keep order:false, stats:pseudo
+    │ └─TableReader_33(Probe)	9980.01	root		data:Selection_32
+    │   └─Selection_32	9980.01	cop[tikv]		not(isnull(explain_cte.t2.c)), not(isnull(explain_cte.t2.d))
+    │     └─TableFullScan_31	10000.00	cop[tikv]	table:t2	keep order:false, stats:pseudo
+    └─TableReader_36(Probe)	9990.00	root		data:Selection_35
+      └─Selection_35	9990.00	cop[tikv]		not(isnull(explain_cte.t1.a))
+        └─TableFullScan_34	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
 delete from  t1 using t1 inner join (select t2.c from t2 inner join (with temp as (select e from t3 where t3.f = 1234) select e from temp) tt on t2.d = tt.e) t on t1.a = t.c;
 select * from t1;
 a	b
diff --git a/executor/index_advise_test.go b/executor/index_advise_test.go
@@ -101,23 +101,22 @@ and b.txn_accno = a.new_accno;`
 	rows := [][]interface{}{
 		{"Update_8"},
 		{"└─IndexJoin_14"},
-		{"  ├─TableReader_25(Build)"},
-		{"  │ └─Selection_24"},
-		{"  │   └─TableFullScan_23"},
+		{"  ├─TableReader_23(Build)"},
+		{"  │ └─Selection_22"},
+		{"  │   └─TableFullScan_21"},
 		{"  └─IndexReader_12(Probe)"},
-		{"    └─Selection_11"},
-		{"      └─IndexRangeScan_10"},
+		{"    └─IndexRangeScan_11"},
 	}
 	tk.MustExec("set @@session.tidb_enable_inl_join_inner_multi_pattern='ON'")
 	tk.MustQuery("explain "+sql).CheckAt([]int{0}, rows)
 	rows = [][]interface{}{
 		{"Update_8"},
-		{"└─HashJoin_10"},
-		{"  ├─IndexReader_17(Build)"},
-		{"  │ └─IndexRangeScan_16"},
-		{"  └─TableReader_14(Probe)"},
-		{"    └─Selection_13"},
-		{"      └─TableFullScan_12"},
+		{"└─HashJoin_12"},
+		{"  ├─TableReader_15(Build)"},
+		{"  │ └─Selection_14"},
+		{"  │   └─TableFullScan_13"},
+		{"  └─IndexReader_18(Probe)"},
+		{"    └─IndexRangeScan_17"},
 	}
 	tk.MustExec("set @@session.tidb_enable_inl_join_inner_multi_pattern='OFF'")
 	tk.MustQuery("explain "+sql).CheckAt([]int{0}, rows)

diff --git a/expression/constant_propagation.go b/expression/constant_propagation.go
@@ -69,7 +69,8 @@ func (s *basePropConstSolver) tryToUpdateEQList(col *Column, con *Constant) (boo
 	return true, false
 }
 
-func validEqualCondHelper(ctx sessionctx.Context, eq *ScalarFunction, colIsLeft bool) (*Column, *Constant) {
+// ValidCompareConstantPredicateHelper checks if the predicate is a compare constant predicate, like "Column xxx Constant"
+func ValidCompareConstantPredicateHelper(eq *ScalarFunction, colIsLeft bool) (*Column, *Constant) {
 	var col *Column
 	var con *Constant
 	colOk := false
@@ -97,14 +98,14 @@ func validEqualCondHelper(ctx sessionctx.Context, eq *ScalarFunction, colIsLeft
 }
 
 // validEqualCond checks if the cond is an expression like [column eq constant].
-func validEqualCond(ctx sessionctx.Context, cond Expression) (*Column, *Constant) {
+func validEqualCond(cond Expression) (*Column, *Constant) {
 	if eq, ok := cond.(*ScalarFunction); ok {
 		if eq.FuncName.L != ast.EQ {
 			return nil, nil
 		}
-		col, con := validEqualCondHelper(ctx, eq, true)
+		col, con := ValidCompareConstantPredicateHelper(eq, true)
 		if col == nil {
-			return validEqualCondHelper(ctx, eq, false)
+			return ValidCompareConstantPredicateHelper(eq, false)
 		}
 		return col, con
 	}
@@ -296,7 +297,7 @@ func (s *propConstSolver) pickNewEQConds(visited []bool) (retMapper map[int]*Con
 		if visited[i] {
 			continue
 		}
-		col, con := validEqualCond(s.ctx, cond)
+		col, con := validEqualCond(cond)
 		// Then we check if this CNF item is a false constant. If so, we will set the whole condition to false.
 		var ok bool
 		if col == nil {
@@ -357,6 +358,7 @@ func (s *propConstSolver) solve(conditions []Expression) []Expression {
 }
 
 // PropagateConstant propagate constant values of deterministic predicates in a condition.
+// This is a constant propagation logic for expression list such as ['a=1', 'a=b']
 func PropagateConstant(ctx sessionctx.Context, conditions []Expression) []Expression {
 	return newPropConstSolver().PropagateConstant(ctx, conditions)
 }
@@ -400,7 +402,7 @@ func (s *propOuterJoinConstSolver) pickEQCondsOnOuterCol(retMapper map[int]*Cons
 		if visited[i+condsOffset] {
 			continue
 		}
-		col, con := validEqualCond(s.ctx, cond)
+		col, con := validEqualCond(cond)
 		// Then we check if this CNF item is a false constant. If so, we will set the whole condition to false.
 		var ok bool
 		if col == nil {

diff --git a/planner/cardinality/selectivity_test.go b/planner/cardinality/selectivity_test.go
@@ -1202,33 +1202,29 @@ func TestIndexJoinInnerRowCountUpperBound(t *testing.T) {
 	stat := h.GetTableStats(tblInfo)
 	stat.HistColl = mockStatsTbl.HistColl
 
-	query := "explain format = 'brief' " +
-		"select /*+ inl_join(t2) */ * from (select * from t where t.a < 1) as t1 join t t2 where t2.a = 0 and t1.a = t2.b"
-
-	testKit.MustQuery(query).Check(testkit.Rows(
-		"IndexJoin 1000000.00 root  inner join, inner:IndexLookUp, outer key:test.t.a, inner key:test.t.b, equal cond:eq(test.t.a, test.t.b)",
-		"├─TableReader(Build) 1000.00 root  data:Selection",
-		"│ └─Selection 1000.00 cop[tikv]  lt(test.t.a, 1), not(isnull(test.t.a))",
-		"│   └─TableFullScan 500000.00 cop[tikv] table:t keep order:false, stats:pseudo",
-		"└─IndexLookUp(Probe) 1000000.00 root  ",
-		"  ├─Selection(Build) 500000000.00 cop[tikv]  not(isnull(test.t.b))",
-		"  │ └─IndexRangeScan 500000000.00 cop[tikv] table:t2, index:idx(b) range: decided by [eq(test.t.b, test.t.a)], keep order:false, stats:pseudo",
-		"  └─Selection(Probe) 1000000.00 cop[tikv]  eq(test.t.a, 0)",
-		"    └─TableRowIDScan 500000000.00 cop[tikv] table:t2 keep order:false, stats:pseudo",
-	))
+	var (
+		input  []string
+		output []struct {
+			Query  string
+			Result []string
+		}
+	)
 
-	testKit.MustExec("set @@tidb_opt_fix_control = '44855:ON'")
-	testKit.MustQuery(query).Check(testkit.Rows(
-		"IndexJoin 1000000.00 root  inner join, inner:IndexLookUp, outer key:test.t.a, inner key:test.t.b, equal cond:eq(test.t.a, test.t.b)",
-		"├─TableReader(Build) 1000.00 root  data:Selection",
-		"│ └─Selection 1000.00 cop[tikv]  lt(test.t.a, 1), not(isnull(test.t.a))",
-		"│   └─TableFullScan 500000.00 cop[tikv] table:t keep order:false, stats:pseudo",
-		"└─IndexLookUp(Probe) 1000000.00 root  ",
-		"  ├─Selection(Build) 1000000.00 cop[tikv]  not(isnull(test.t.b))",
-		"  │ └─IndexRangeScan 1000000.00 cop[tikv] table:t2, index:idx(b) range: decided by [eq(test.t.b, test.t.a)], keep order:false, stats:pseudo",
-		"  └─Selection(Probe) 1000000.00 cop[tikv]  eq(test.t.a, 0)",
-		"    └─TableRowIDScan 1000000.00 cop[tikv] table:t2 keep order:false, stats:pseudo",
-	))
+	suiteData := cardinality.GetCardinalitySuiteData()
+	suiteData.LoadTestCases(t, &input, &output)
+	for i := 0; i < len(input); i++ {
+		testdata.OnRecord(func() {
+			output[i].Query = input[i]
+		})
+		if !strings.HasPrefix(input[i], "explain") {
+			testKit.MustExec(input[i])
+			continue
+		}
+		testdata.OnRecord(func() {
+			output[i].Result = testdata.ConvertRowsToStrings(testKit.MustQuery(input[i]).Rows())
+		})
+		testKit.MustQuery(input[i]).Check(testkit.Rows(output[i].Result...))
+	}
 }
 
 func TestOrderingIdxSelectivityThreshold(t *testing.T) {

diff --git a/planner/cardinality/testdata/cardinality_suite_in.json b/planner/cardinality/testdata/cardinality_suite_in.json
@@ -454,5 +454,13 @@
       "select * from t where a = 100 and b = 350",
       "select * from t where a < -1500 and b > 400 and b < 403"
     ]
+  },
+  {
+    "name": "TestIndexJoinInnerRowCountUpperBound",
+    "cases": [
+      "explain format = 'brief' select /*+ inl_join(t2) */ * from (select * from t where t.a < 1) as t1 join t t2 where t2.a = 0 and t1.a = t2.b",
+      "set @@tidb_opt_fix_control = '44855:ON'",
+      "explain format = 'brief' select /*+ inl_join(t2) */ * from (select * from t where t.a < 1) as t1 join t t2 where t2.a = 0 and t1.a = t2.b"
+    ]
   }
 ]
diff --git a/planner/cardinality/testdata/cardinality_suite_out.json b/planner/cardinality/testdata/cardinality_suite_out.json
@@ -4630,5 +4630,44 @@
         ]
       }
     ]
+  },
+  {
+    "Name": "TestIndexJoinInnerRowCountUpperBound",
+    "Cases": [
+      {
+        "Query": "explain format = 'brief' select /*+ inl_join(t2) */ * from (select * from t where t.a < 1) as t1 join t t2 where t2.a = 0 and t1.a = t2.b",
+        "Result": [
+          "Projection 2000.00 root  test.t.a, test.t.b, test.t.a, test.t.b",
+          "└─IndexJoin 2000.00 root  inner join, inner:IndexLookUp, outer key:test.t.a, inner key:test.t.b, equal cond:eq(test.t.a, test.t.b)",
+          "  ├─TableReader(Build) 1000.00 root  data:Selection",
+          "  │ └─Selection 1000.00 cop[tikv]  lt(test.t.a, 1), not(isnull(test.t.a))",
+          "  │   └─TableFullScan 500000.00 cop[tikv] table:t keep order:false, stats:pseudo",
+          "  └─IndexLookUp(Probe) 2000.00 root  ",
+          "    ├─Selection(Build) 1000000.00 cop[tikv]  lt(test.t.b, 1), not(isnull(test.t.b))",
+          "    │ └─IndexRangeScan 500000000.00 cop[tikv] table:t2, index:idx(b) range: decided by [eq(test.t.b, test.t.a)], keep order:false, stats:pseudo",
+          "    └─Selection(Probe) 2000.00 cop[tikv]  eq(test.t.a, 0)",
+          "      └─TableRowIDScan 1000000.00 cop[tikv] table:t2 keep order:false, stats:pseudo"
+        ]
+      },
+      {
+        "Query": "set @@tidb_opt_fix_control = '44855:ON'",
+        "Result": null
+      },
+      {
+        "Query": "explain format = 'brief' select /*+ inl_join(t2) */ * from (select * from t where t.a < 1) as t1 join t t2 where t2.a = 0 and t1.a = t2.b",
+        "Result": [
+          "Projection 2000.00 root  test.t.a, test.t.b, test.t.a, test.t.b",
+          "└─IndexJoin 2000.00 root  inner join, inner:IndexLookUp, outer key:test.t.a, inner key:test.t.b, equal cond:eq(test.t.a, test.t.b)",
+          "  ├─TableReader(Build) 1000.00 root  data:Selection",
+          "  │ └─Selection 1000.00 cop[tikv]  lt(test.t.a, 1), not(isnull(test.t.a))",
+          "  │   └─TableFullScan 500000.00 cop[tikv] table:t keep order:false, stats:pseudo",
+          "  └─IndexLookUp(Probe) 2000.00 root  ",
+          "    ├─Selection(Build) 1000000.00 cop[tikv]  lt(test.t.b, 1), not(isnull(test.t.b))",
+          "    │ └─IndexRangeScan 1000000.00 cop[tikv] table:t2, index:idx(b) range: decided by [eq(test.t.b, test.t.a)], keep order:false, stats:pseudo",
+          "    └─Selection(Probe) 2000.00 cop[tikv]  eq(test.t.a, 0)",
+          "      └─TableRowIDScan 1000000.00 cop[tikv] table:t2 keep order:false, stats:pseudo"
+        ]
+      }
+    ]
   }
 ]
diff --git a/planner/core/BUILD.bazel b/planner/core/BUILD.bazel
@@ -50,6 +50,7 @@ go_library(
         "rule_aggregation_skew_rewrite.go",
         "rule_build_key_info.go",
         "rule_column_pruning.go",
+        "rule_constant_propagation.go",
         "rule_decorrelate.go",
         "rule_derive_topn_from_window.go",
         "rule_eliminate_projection.go",
@@ -221,6 +222,7 @@ go_test(
         "planbuilder_test.go",
         "point_get_plan_test.go",
         "preprocess_test.go",
+        "rule_constant_propagation_test.go",
         "rule_generate_column_substitute_test.go",
         "rule_join_reorder_dp_test.go",
         "rule_join_reorder_test.go",