planner: correct estimated row count for inner plan of index join (#10015)

Author: eurekaka

cmd/explaintest/r/explain_easy.result

@@ -61,8 +61,8 @@ explain select count(b.c2) from t1 a, t2 b where a.c1 = b.c2 group by a.c1;
 id	count	task	operator info
 Projection_11	9990.00	root	cast(join_agg_0)
 └─IndexJoin_14	9990.00	root	inner join, inner:TableReader_13, outer key:b.c2, inner key:a.c1
-  ├─TableReader_13	10.00	root	data:TableScan_12
-  │ └─TableScan_12	10.00	cop	table:a, range: decided by [b.c2], keep order:false, stats:pseudo
+  ├─TableReader_13	1.00	root	data:TableScan_12
+  │ └─TableScan_12	1.00	cop	table:a, range: decided by [b.c2], keep order:false, stats:pseudo
   └─HashAgg_22	7992.00	root	group by:col_2, funcs:count(col_0), firstrow(col_1)
     └─TableReader_23	7992.00	root	data:HashAgg_17
       └─HashAgg_17	7992.00	cop	group by:b.c2, funcs:count(b.c2), firstrow(b.c2)
@@ -99,8 +99,8 @@ explain select c1 from t1 where c1 in (select c2 from t2);
 id	count	task	operator info
 Projection_9	9990.00	root	test.t1.c1
 └─IndexJoin_12	9990.00	root	inner join, inner:TableReader_11, outer key:test.t2.c2, inner key:test.t1.c1
-  ├─TableReader_11	10.00	root	data:TableScan_10
-  │ └─TableScan_10	10.00	cop	table:t1, range: decided by [test.t2.c2], keep order:false, stats:pseudo
+  ├─TableReader_11	1.00	root	data:TableScan_10
+  │ └─TableScan_10	1.00	cop	table:t1, range: decided by [test.t2.c2], keep order:false, stats:pseudo
   └─HashAgg_20	7992.00	root	group by:col_1, funcs:firstrow(col_0)
     └─TableReader_21	7992.00	root	data:HashAgg_15
       └─HashAgg_15	7992.00	cop	group by:test.t2.c2, funcs:firstrow(test.t2.c2)
@@ -291,9 +291,9 @@ Projection_11	10000.00	root	9_aux_0
     └─IndexJoin_44	12.50	root	inner join, inner:TableReader_43, outer key:s.a, inner key:t1.a
       ├─TableReader_37	1.00	root	data:TableScan_36
       │ └─TableScan_36	1.00	cop	table:s, range: decided by [eq(s.a, test.t.a)], keep order:false, stats:pseudo
-      └─TableReader_43	10.00	root	data:Selection_42
-        └─Selection_42	10.00	cop	eq(t1.a, test.t.a)
-          └─TableScan_41	10.00	cop	table:t1, range: decided by [s.a], keep order:false, stats:pseudo
+      └─TableReader_43	0.80	root	data:Selection_42
+        └─Selection_42	0.80	cop	eq(t1.a, test.t.a)
+          └─TableScan_41	1.00	cop	table:t1, range: decided by [s.a], keep order:false, stats:pseudo
 explain select t.c in (select count(*) from t s use index(idx), t t1 where s.b = t.a and s.a = t1.a) from t;
 id	count	task	operator info
 Projection_11	10000.00	root	9_aux_0
@@ -304,8 +304,8 @@ Projection_11	10000.00	root	9_aux_0
     └─IndexJoin_32	12.50	root	inner join, inner:TableReader_31, outer key:s.a, inner key:t1.a
       ├─IndexReader_27	10.00	root	index:IndexScan_26
       │ └─IndexScan_26	10.00	cop	table:s, index:b, range: decided by [eq(s.b, test.t.a)], keep order:false, stats:pseudo
-      └─TableReader_31	10.00	root	data:TableScan_30
-        └─TableScan_30	10.00	cop	table:t1, range: decided by [s.a], keep order:false, stats:pseudo
+      └─TableReader_31	1.00	root	data:TableScan_30
+        └─TableScan_30	1.00	cop	table:t1, range: decided by [s.a], keep order:false, stats:pseudo
 explain select t.c in (select count(*) from t s use index(idx), t t1 where s.b = t.a and s.c = t1.a) from t;
 id	count	task	operator info
 Projection_11	10000.00	root	9_aux_0
@@ -318,8 +318,8 @@ Projection_11	10000.00	root	9_aux_0
       │ ├─IndexScan_26	10.00	cop	table:s, index:b, range: decided by [eq(s.b, test.t.a)], keep order:false, stats:pseudo
       │ └─Selection_28	9.99	cop	not(isnull(s.c))
       │   └─TableScan_27	10.00	cop	table:t, keep order:false, stats:pseudo
-      └─TableReader_33	10.00	root	data:TableScan_32
-        └─TableScan_32	10.00	cop	table:t1, range: decided by [s.c], keep order:false, stats:pseudo
+      └─TableReader_33	1.00	root	data:TableScan_32
+        └─TableScan_32	1.00	cop	table:t1, range: decided by [s.c], keep order:false, stats:pseudo
 drop table if exists t;
 create table t(a int unsigned);
 explain select t.a = '123455' from t;

cmd/explaintest/r/tpch.result

@@ -360,8 +360,8 @@ Sort_23	5.00	root	revenue:desc
         │ │ │   └─TableScan_56	500000.00	cop	table:supplier, range:[-inf,+inf], keep order:false
         │ │ └─TableReader_59	300005811.00	root	data:TableScan_58
         │ │   └─TableScan_58	300005811.00	cop	table:lineitem, range:[-inf,+inf], keep order:false
-        │ └─TableReader_37	11822812.50	root	data:Selection_36
-        │   └─Selection_36	11822812.50	cop	ge(tpch.orders.o_orderdate, 1994-01-01 00:00:00.000000), lt(tpch.orders.o_orderdate, 1995-01-01)
+        │ └─TableReader_37	0.80	root	data:Selection_36
+        │   └─Selection_36	0.80	cop	ge(tpch.orders.o_orderdate, 1994-01-01 00:00:00.000000), lt(tpch.orders.o_orderdate, 1995-01-01)
         │     └─TableScan_35	1.00	cop	table:orders, range: decided by [tpch.lineitem.l_orderkey], keep order:false
         └─TableReader_30	1.00	root	data:TableScan_29
           └─TableScan_29	1.00	cop	table:customer, range: decided by [tpch.supplier.s_nationkey tpch.orders.o_custkey], keep order:false
@@ -1096,8 +1096,8 @@ StreamAgg_13	1.00	root	funcs:sum(col_0)
     ├─TableReader_22	6286493.79	root	data:Selection_21
     │ └─Selection_21	6286493.79	cop	eq(tpch.lineitem.l_shipinstruct, "DELIVER IN PERSON"), in(tpch.lineitem.l_shipmode, "AIR", "AIR REG"), or(and(ge(tpch.lineitem.l_quantity, 4), le(tpch.lineitem.l_quantity, 14)), or(and(ge(tpch.lineitem.l_quantity, 18), le(tpch.lineitem.l_quantity, 28)), and(ge(tpch.lineitem.l_quantity, 29), le(tpch.lineitem.l_quantity, 39))))
     │   └─TableScan_20	300005811.00	cop	table:lineitem, range:[-inf,+inf], keep order:false
-    └─TableReader_28	8000000.00	root	data:Selection_27
-      └─Selection_27	8000000.00	cop	ge(tpch.part.p_size, 1), or(and(eq(tpch.part.p_brand, "Brand#52"), and(in(tpch.part.p_container, "SM CASE", "SM BOX", "SM PACK", "SM PKG"), le(tpch.part.p_size, 5))), or(and(eq(tpch.part.p_brand, "Brand#11"), and(in(tpch.part.p_container, "MED BAG", "MED BOX", "MED PKG", "MED PACK"), le(tpch.part.p_size, 10))), and(eq(tpch.part.p_brand, "Brand#51"), and(in(tpch.part.p_container, "LG CASE", "LG BOX", "LG PACK", "LG PKG"), le(tpch.part.p_size, 15)))))
+    └─TableReader_28	0.80	root	data:Selection_27
+      └─Selection_27	0.80	cop	ge(tpch.part.p_size, 1), or(and(eq(tpch.part.p_brand, "Brand#52"), and(in(tpch.part.p_container, "SM CASE", "SM BOX", "SM PACK", "SM PKG"), le(tpch.part.p_size, 5))), or(and(eq(tpch.part.p_brand, "Brand#11"), and(in(tpch.part.p_container, "MED BAG", "MED BOX", "MED PKG", "MED PACK"), le(tpch.part.p_size, 10))), and(eq(tpch.part.p_brand, "Brand#51"), and(in(tpch.part.p_container, "LG CASE", "LG BOX", "LG PACK", "LG PKG"), le(tpch.part.p_size, 15)))))
         └─TableScan_26	1.00	cop	table:part, range: decided by [tpch.lineitem.l_partkey], keep order:false
 /*
 Q20 Potential Part Promotion Query
@@ -1235,8 +1235,8 @@ Projection_25	1.00	root	tpch.supplier.s_name, 17_col_0
       │ │ │ └─TableReader_80	240004648.80	root	data:Selection_79
       │ │ │   └─Selection_79	240004648.80	cop	gt(l1.l_receiptdate, l1.l_commitdate)
       │ │ │     └─TableScan_78	300005811.00	cop	table:l1, range:[-inf,+inf], keep order:false
-      │ │ └─TableReader_61	36517371.00	root	data:Selection_60
-      │ │   └─Selection_60	36517371.00	cop	eq(tpch.orders.o_orderstatus, "F")
+      │ │ └─TableReader_61	0.80	root	data:Selection_60
+      │ │   └─Selection_60	0.80	cop	eq(tpch.orders.o_orderstatus, "F")
       │ │     └─TableScan_59	1.00	cop	table:orders, range: decided by [l1.l_orderkey], keep order:false
       │ └─IndexLookUp_55	1.00	root	
       │   ├─IndexScan_53	1.00	cop	table:l2, index:L_ORDERKEY, L_LINENUMBER, range: decided by [l1.l_orderkey], keep order:false

executor/index_lookup_join_test.go

@@ -53,9 +53,9 @@ func (s *testSuite1) TestIndexJoinUnionScan(c *C) {
 		"│ └─TableReader_15 9990.00 root data:Selection_14",
 		"│   └─Selection_14 9990.00 cop not(isnull(test.t1.a))",
 		"│     └─TableScan_13 10000.00 cop table:t1, range:[-inf,+inf], keep order:false, stats:pseudo",
-		"└─UnionScan_10 10.00 root ",
-		"  └─TableReader_9 10.00 root data:TableScan_8",
-		"    └─TableScan_8 10.00 cop table:t2, range: decided by [test.t1.a], keep order:false, stats:pseudo",
+		"└─UnionScan_10 1.00 root ",
+		"  └─TableReader_9 1.00 root data:TableScan_8",
+		"    └─TableScan_8 1.00 cop table:t2, range: decided by [test.t1.a], keep order:false, stats:pseudo",
 	))
 	tk.MustQuery("select /*+ TIDB_INLJ(t1, t2)*/ * from t1 join t2 on t1.a = t2.id").Check(testkit.Rows(
 		"2 2 2 2 2",

planner/core/exhaust_physical_plans.go

@@ -461,22 +461,15 @@ func (p *LogicalJoin) constructInnerTableScan(ds *DataSource, pk *expression.Col
 	}.Init(ds.ctx)
 	ts.SetSchema(ds.schema)
 
-	var rowCount float64
-	pkHist, ok := ds.statisticTable.Columns[pk.ID]
-	if ok && !ds.statisticTable.Pseudo {
-		rowCount = pkHist.AvgCountPerNotNullValue(ds.statisticTable.Count)
-	} else {
-		rowCount = ds.statisticTable.PseudoAvgCountPerValue()
-	}
-
-	ts.stats = property.NewSimpleStats(rowCount)
+	ts.stats = property.NewSimpleStats(1)
 	ts.stats.UsePseudoStats = ds.statisticTable.Pseudo
 
 	copTask := &copTask{
 		tablePlan:         ts,
 		indexPlanFinished: true,
 	}
-	ts.addPushedDownSelection(copTask, ds.stats)
+	selStats := ts.stats.Scale(selectionFactor)
+	ts.addPushedDownSelection(copTask, selStats)
 	t := finishCopTask(ds.ctx, copTask)
 	reader := t.plan()
 	return p.constructInnerUnionScan(us, reader)