plan: make hashjoin's cost more reasonable. (pingcap#4022)

rauno-palosaari · Aug 4, 2017 · a5b26bb · a5b26bb
1 parent de14330
commit a5b26bb
Show file tree

Hide file tree

Showing 3 changed files with 21 additions and 14 deletions.
diff --git a/plan/dag_plan_test.go b/plan/dag_plan_test.go
@@ -191,13 +191,21 @@ func (s *testPlanSuite) TestDAGPlanBuilderJoin(c *C) {
 			sql:  "select * from t t1 join t t2 on t1.a = t2.c_str",
 			best: "LeftHashJoin{TableReader(Table(t))->Projection->TableReader(Table(t))->Projection}(cast(t1.a),cast(t2.c_str))->Projection",
 		},
+		{
+			sql:  "select * from t t1 join t t2 on t1.b = t2.a",
+			best: "LeftHashJoin{TableReader(Table(t))->TableReader(Table(t))}(t1.b,t2.a)",
+		},
 		{
 			sql:  "select * from t t1 join t t2 on t1.a = t2.a join t t3 on t1.a = t3.a",
 			best: "MergeJoin{MergeJoin{TableReader(Table(t))->TableReader(Table(t))}(t1.a,t2.a)->TableReader(Table(t))}(t1.a,t3.a)",
 		},
+		{
+			sql:  "select * from t t1 join t t2 on t1.a = t2.a join t t3 on t1.b = t3.a",
+			best: "LeftHashJoin{MergeJoin{TableReader(Table(t))->TableReader(Table(t))}(t1.a,t2.a)->TableReader(Table(t))}(t1.b,t3.a)",
+		},
 		{
 			sql:  "select * from t t1 join t t2 on t1.b = t2.a order by t1.a",
-			best: "IndexJoin{TableReader(Table(t))->TableReader(Table(t))}(t1.b,t2.a)->Sort",
+			best: "LeftHashJoin{TableReader(Table(t))->TableReader(Table(t))}(t1.b,t2.a)->Sort",
 		},
 		{
 			sql:  "select * from t t1 join t t2 on t1.b = t2.a order by t1.a limit 1",
@@ -217,7 +225,7 @@ func (s *testPlanSuite) TestDAGPlanBuilderJoin(c *C) {
 		},
 		{
 			sql:  "select * from t t1 join t t2 on t1.a = t2.a join t t3 on t1.a = t3.a and t1.b = 1 and t3.c = 1",
-			best: "IndexJoin{IndexJoin{TableReader(Table(t)->Sel([eq(t1.b, 1)]))->TableReader(Table(t))}(t1.a,t2.a)->TableReader(Table(t)->Sel([eq(t3.c, 1)]))}(t1.a,t3.a)",
+			best: "LeftHashJoin{IndexJoin{TableReader(Table(t)->Sel([eq(t1.b, 1)]))->TableReader(Table(t))}(t1.a,t2.a)->IndexLookUp(Index(t.c_d_e)[[1,1]], Table(t))}(t1.a,t3.a)",
 		},
 		{
 			sql:  "select * from t where t.c in (select b from t s where s.a = t.a)",
@@ -231,7 +239,7 @@ func (s *testPlanSuite) TestDAGPlanBuilderJoin(c *C) {
 		// Merge Join will no longer enforce a sort. If a hint doesn't take effect, we will choose other types of join.
 		{
 			sql:  "select /*+ TIDB_SMJ(t1,t2)*/ * from t t1, t t2 where t1.a = t2.b",
-			best: "IndexJoin{TableReader(Table(t))->TableReader(Table(t))}(t2.b,t1.a)->Projection",
+			best: "LeftHashJoin{TableReader(Table(t))->TableReader(Table(t))}(t1.a,t2.b)",
 		},
 		// Test Single Merge Join + Sort.
 		{
@@ -301,7 +309,7 @@ func (s *testPlanSuite) TestDAGPlanBuilderJoin(c *C) {
 		// Test Index Join failed.
 		{
 			sql:  "select /*+ TIDB_INLJ(t1) */ * from t t1 right outer join t t2 on t1.a = t2.b",
-			best: "IndexJoin{TableReader(Table(t))->TableReader(Table(t))}(t2.b,t1.a)->Projection",
+			best: "RightHashJoin{TableReader(Table(t))->TableReader(Table(t))}(t1.a,t2.b)",
 		},
 	}
 	for _, tt := range tests {

diff --git a/plan/explain_test.go b/plan/explain_test.go
@@ -98,11 +98,10 @@ func (s *testExplainSuite) TestExplain(c *C) {
 			"select * from t1 left join t2 on t1.c2 = t2.c1 where t1.c1 > 1",
 			[]string{
 				"TableScan_22  cop table:t1, range:[2,+inf), keep order:false",
-				"TableReader_23 IndexJoin_7 root data:TableScan_22",
-				"IndexScan_33  cop table:t2, index:c1, range:[<nil>,+inf], out of order:false",
-				"TableScan_34  cop table:t2, keep order:false",
-				"IndexLookUp_35 IndexJoin_7 root index:IndexScan_33, table:TableScan_34",
-				"IndexJoin_7  root outer:TableReader_23, outer key:test.t1.c2, inner key:test.t2.c1",
+				"TableReader_23 HashLeftJoin_8 root data:TableScan_22",
+				"TableScan_37  cop table:t2, range:(-inf,+inf), keep order:false",
+				"TableReader_38 HashLeftJoin_8 root data:TableScan_37",
+				"HashLeftJoin_8  root left outer join, small:TableReader_38, equal:[eq(test.t1.c2, test.t2.c1)]",
 			},
 		},
 		{
@@ -125,13 +124,13 @@ func (s *testExplainSuite) TestExplain(c *C) {
 		{
 			"select count(b.c2) from t1 a, t2 b where a.c1 = b.c2 group by a.c1",
 			[]string{
+				"TableScan_25  cop table:a, range:(-inf,+inf), keep order:false",
+				"TableReader_26 HashLeftJoin_10 root data:TableScan_25",
 				"TableScan_17 HashAgg_16 cop table:b, range:(-inf,+inf), keep order:false",
 				"HashAgg_16  cop type:complete, group by:b.c2, funcs:count(b.c2), firstrow(b.c2)",
 				"TableReader_19 HashAgg_18 root data:HashAgg_16",
-				"HashAgg_18 IndexJoin_9 root type:final, group by:, funcs:count(col_0), firstrow(col_1)",
-				"TableScan_12  cop table:a, range:(-inf,+inf), keep order:true",
-				"TableReader_22 IndexJoin_9 root data:TableScan_12",
-				"IndexJoin_9 Projection_8 root outer:TableReader_22, outer key:b.c2, inner key:a.c1",
+				"HashAgg_18 HashLeftJoin_10 root type:final, group by:, funcs:count(col_0), firstrow(col_1)",
+				"HashLeftJoin_10 Projection_8 root inner join, small:HashAgg_18, equal:[eq(a.c1, b.c2)]",
 				"Projection_8  root cast(join_agg_0)",
 			},
 		},

diff --git a/plan/task.go b/plan/task.go
@@ -166,7 +166,7 @@ func (p *PhysicalHashJoin) getCost(lCnt, rCnt float64) float64 {
 	if smallTableCnt <= 1 {
 		smallTableCnt = 1
 	}
-	return (lCnt + rCnt) * (1 + math.Log2(smallTableCnt))
+	return (lCnt + rCnt) * (1 + math.Log2(smallTableCnt)/float64(p.Concurrency))
 }
 
 func (p *PhysicalHashJoin) attach2Task(tasks ...task) task {