Skip to content

Commit

Permalink
planner: build anti semi join for NOT EXISTS (#7842) (#11291)
Browse files Browse the repository at this point in the history
  • Loading branch information
@eurekaka
eurekaka authored Jul 18, 2019
1 parent 89032d7 commit 6626b49
Show file tree
Hide file tree
Showing 5 changed files with 37 additions and 71 deletions.
64 changes: 31 additions & 33 deletions cmd/explaintest/r/tpch.result
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1223,31 +1223,30 @@ id count task operator info
Projection_25 100.00 root tpch.supplier.s_name, 17_col_0
└─TopN_28 100.00 root 17_col_0:desc, tpch.supplier.s_name:asc, offset:0, count:100
└─HashAgg_31 320000.00 root group by:tpch.supplier.s_name, funcs:count(1), firstrow(tpch.supplier.s_name)
└─Selection_32 3786715.90 root not(16_aux_0)
└─IndexJoin_38 4733394.87 root left outer semi join, inner:IndexLookUp_37, outer key:tpch.l1.l_orderkey, inner key:tpch.l3.l_orderkey, other cond:ne(tpch.l3.l_suppkey, tpch.l1.l_suppkey)
├─IndexJoin_82 4733394.87 root semi join, inner:IndexLookUp_81, outer key:tpch.l1.l_orderkey, inner key:tpch.l2.l_orderkey, other cond:ne(tpch.l2.l_suppkey, tpch.l1.l_suppkey), ne(tpch.l2.l_suppkey, tpch.supplier.s_suppkey)
│ ├─HashLeftJoin_88 5916743.59 root inner join, inner:TableReader_117, equal:[eq(tpch.supplier.s_nationkey, tpch.nation.n_nationkey)]
│ │ ├─HashLeftJoin_93 147918589.81 root inner join, inner:TableReader_114, equal:[eq(tpch.l1.l_suppkey, tpch.supplier.s_suppkey)]
│ │ │ ├─IndexJoin_100 147918589.81 root inner join, inner:IndexLookUp_99, outer key:tpch.orders.o_orderkey, inner key:tpch.l1.l_orderkey
│ │ │ │ ├─TableReader_109 36517371.00 root data:Selection_108
│ │ │ │ │ └─Selection_108 36517371.00 cop eq(tpch.orders.o_orderstatus, "F")
│ │ │ │ │ └─TableScan_107 75000000.00 cop table:orders, range:[-inf,+inf], keep order:false
│ │ │ │ └─IndexLookUp_99 0.80 root
│ │ │ │ ├─IndexScan_96 1.00 cop table:l1, index:L_ORDERKEY, L_LINENUMBER, range: decided by [tpch.orders.o_orderkey], keep order:false
│ │ │ │ └─Selection_98 0.80 cop gt(tpch.l1.l_receiptdate, tpch.l1.l_commitdate)
│ │ │ │ └─TableScan_97 1.00 cop table:lineitem, keep order:false
│ │ │ └─TableReader_114 500000.00 root data:TableScan_113
│ │ │ └─TableScan_113 500000.00 cop table:supplier, range:[-inf,+inf], keep order:false
│ │ └─TableReader_117 1.00 root data:Selection_116
│ │ └─Selection_116 1.00 cop eq(tpch.nation.n_name, "EGYPT")
│ │ └─TableScan_115 25.00 cop table:nation, range:[-inf,+inf], keep order:false
│ └─IndexLookUp_81 1.00 root
│ ├─IndexScan_79 1.00 cop table:l2, index:L_ORDERKEY, L_LINENUMBER, range: decided by [tpch.l1.l_orderkey], keep order:false
│ └─TableScan_80 1.00 cop table:lineitem, keep order:false
└─IndexLookUp_37 0.80 root
├─IndexScan_34 1.00 cop table:l3, index:L_ORDERKEY, L_LINENUMBER, range: decided by [tpch.l1.l_orderkey], keep order:false
└─Selection_36 0.80 cop gt(tpch.l3.l_receiptdate, tpch.l3.l_commitdate)
└─TableScan_35 1.00 cop table:lineitem, keep order:false
└─IndexJoin_37 3786715.90 root anti semi join, inner:IndexLookUp_36, outer key:tpch.l1.l_orderkey, inner key:tpch.l3.l_orderkey, other cond:ne(tpch.l3.l_suppkey, tpch.l1.l_suppkey), ne(tpch.l3.l_suppkey, tpch.supplier.s_suppkey)
├─IndexJoin_81 4733394.87 root semi join, inner:IndexLookUp_80, outer key:tpch.l1.l_orderkey, inner key:tpch.l2.l_orderkey, other cond:ne(tpch.l2.l_suppkey, tpch.l1.l_suppkey), ne(tpch.l2.l_suppkey, tpch.supplier.s_suppkey)
│ ├─HashLeftJoin_87 5916743.59 root inner join, inner:TableReader_116, equal:[eq(tpch.supplier.s_nationkey, tpch.nation.n_nationkey)]
│ │ ├─HashLeftJoin_92 147918589.81 root inner join, inner:TableReader_113, equal:[eq(tpch.l1.l_suppkey, tpch.supplier.s_suppkey)]
│ │ │ ├─IndexJoin_99 147918589.81 root inner join, inner:IndexLookUp_98, outer key:tpch.orders.o_orderkey, inner key:tpch.l1.l_orderkey
│ │ │ │ ├─TableReader_108 36517371.00 root data:Selection_107
│ │ │ │ │ └─Selection_107 36517371.00 cop eq(tpch.orders.o_orderstatus, "F")
│ │ │ │ │ └─TableScan_106 75000000.00 cop table:orders, range:[-inf,+inf], keep order:false
│ │ │ │ └─IndexLookUp_98 0.80 root
│ │ │ │ ├─IndexScan_95 1.00 cop table:l1, index:L_ORDERKEY, L_LINENUMBER, range: decided by [tpch.orders.o_orderkey], keep order:false
│ │ │ │ └─Selection_97 0.80 cop gt(tpch.l1.l_receiptdate, tpch.l1.l_commitdate)
│ │ │ │ └─TableScan_96 1.00 cop table:lineitem, keep order:false
│ │ │ └─TableReader_113 500000.00 root data:TableScan_112
│ │ │ └─TableScan_112 500000.00 cop table:supplier, range:[-inf,+inf], keep order:false
│ │ └─TableReader_116 1.00 root data:Selection_115
│ │ └─Selection_115 1.00 cop eq(tpch.nation.n_name, "EGYPT")
│ │ └─TableScan_114 25.00 cop table:nation, range:[-inf,+inf], keep order:false
│ └─IndexLookUp_80 1.00 root
│ ├─IndexScan_78 1.00 cop table:l2, index:L_ORDERKEY, L_LINENUMBER, range: decided by [tpch.l1.l_orderkey], keep order:false
│ └─TableScan_79 1.00 cop table:lineitem, keep order:false
└─IndexLookUp_36 0.80 root
├─IndexScan_33 1.00 cop table:l3, index:L_ORDERKEY, L_LINENUMBER, range: decided by [tpch.l1.l_orderkey], keep order:false
└─Selection_35 0.80 cop gt(tpch.l3.l_receiptdate, tpch.l3.l_commitdate)
└─TableScan_34 1.00 cop table:lineitem, keep order:false
/*
Q22 Global Sales Opportunity Query
The Global Sales Opportunity Query identifies geographies where there are customers who may be likely to make a
Expand Down Expand Up @@ -1299,11 +1298,10 @@ Sort_32 1.00 root custsale.cntrycode:asc
└─Projection_34 1.00 root custsale.cntrycode, 28_col_0, 28_col_1
└─HashAgg_37 1.00 root group by:custsale.cntrycode, funcs:count(1), sum(tpch.custsale.c_acctbal), firstrow(custsale.cntrycode)
└─Projection_38 0.00 root substring(tpch.customer.c_phone, 1, 2), tpch.customer.c_acctbal
└─Selection_39 0.00 root not(26_aux_0)
└─HashLeftJoin_40 0.00 root left outer semi join, inner:TableReader_46, equal:[eq(tpch.customer.c_custkey, tpch.orders.o_custkey)]
├─Selection_41 0.00 root in(substring(tpch.customer.c_phone, 1, 2), "20", "40", "22", "30", "39", "42", "21")
│ └─TableReader_44 0.00 root data:Selection_43
│ └─Selection_43 0.00 cop gt(tpch.customer.c_acctbal, NULL)
│ └─TableScan_42 7500000.00 cop table:customer, range:[-inf,+inf], keep order:false
└─TableReader_46 75000000.00 root data:TableScan_45
└─TableScan_45 75000000.00 cop table:orders, range:[-inf,+inf], keep order:false
└─HashLeftJoin_39 0.00 root anti semi join, inner:TableReader_45, equal:[eq(tpch.customer.c_custkey, tpch.orders.o_custkey)]
├─Selection_40 0.00 root in(substring(tpch.customer.c_phone, 1, 2), "20", "40", "22", "30", "39", "42", "21")
│ └─TableReader_43 0.00 root data:Selection_42
│ └─Selection_42 0.00 cop gt(tpch.customer.c_acctbal, NULL)
│ └─TableScan_41 7500000.00 cop table:customer, range:[-inf,+inf], keep order:false
└─TableReader_45 75000000.00 root data:TableScan_44
└─TableScan_44 75000000.00 cop table:orders, range:[-inf,+inf], keep order:false
34 changes: 0 additions & 34 deletions executor/join_test.go
View file
Original file line number Diff line number Diff line change
Expand Up @@ -963,40 +963,6 @@ func (s *testSuite) TestHashJoin(c *C) {
c.Assert(outerExecInfo[len(outerExecInfo)-1:], Equals, "1")
innerExecInfo := row[3][4].(string)
c.Assert(innerExecInfo[len(innerExecInfo)-1:], Equals, "0")

tk.MustExec("insert into t2 select * from t1;")
tk.MustExec("delete from t1;")
tk.MustQuery("select count(*) from t1").Check(testkit.Rows("0"))
tk.MustQuery("select count(*) from t2").Check(testkit.Rows("5"))
result = tk.MustQuery("explain analyze select /*+ TIDB_HJ(t1, t2) */ * from t1 where not exists (select a from t2 where t1.a = t2.a);")
// id count task operator info execution info |
// Projection_8 4.00 root test.t1.a, test.t1.b time:193.08µs, loops:1, rows:0 |
// └─Selection_9 4.00 root not(6_aux_0) time:146.95µs, loops:1, rows:0 |
// └─HashLeftJoin_10 5.00 root left outer semi join, inner:TableReader_14, equal:[eq(test.t1.a, test.t2.a)] time:144.293µs, loops:1, rows:0 |
// ├─TableReader_12 5.00 root data:TableScan_11 time:26.27µs, loops:1, rows:0 |
// │ └─TableScan_11 5.00 cop table:t1, range:[-inf,+inf], keep order:false, stats:pseudo |
// └─TableReader_14 5.00 root data:TableScan_13 time:0s, loops:0, rows:0 |
// └─TableScan_13 5.00 cop table:t2, range:[-inf,+inf], keep order:false, stats:pseudo |
row = result.Rows()
c.Assert(len(row), Equals, 7)
outerExecInfo = row[3][4].(string)
c.Assert(outerExecInfo[len(outerExecInfo)-1:], Equals, "0")
innerExecInfo = row[5][4].(string)
c.Assert(innerExecInfo[len(innerExecInfo)-1:], LessEqual, "5")

result = tk.MustQuery("explain analyze select /*+ TIDB_HJ(t1, t2) */ * from t1 left outer join t2 on t1.a = t2.a;")
// id count task operator info execution info
// HashLeftJoin_6 12500.00 root left outer join, inner:TableReader_10, equal:[eq(test.t1.a, test.t2.a)] time:502.553µs, loops:1, rows:0
// ├─TableReader_8 10000.00 root data:TableScan_7 time:27.302µs, loops:1, rows:0
// │ └─TableScan_7 10000.00 cop table:t1, range:[-inf,+inf], keep order:false, stats:pseudo
// └─TableReader_10 10000.00 root data:TableScan_9 time:0s, loops:0, rows:0
// └─TableScan_9 10000.00 cop table:t2, range:[-inf,+inf], keep order:false, stats:pseudo
row = result.Rows()
c.Assert(len(row), Equals, 5)
outerExecInfo = row[1][4].(string)
c.Assert(outerExecInfo[len(outerExecInfo)-1:], Equals, "0")
innerExecInfo = row[3][4].(string)
c.Assert(innerExecInfo[len(innerExecInfo)-1:], LessEqual, "5")
}

func (s *testSuite) TestJoinDifferentDecimals(c *C) {
Expand Down
2 changes: 2 additions & 0 deletions go.mod
View file
Original file line number Diff line number Diff line change
Expand Up @@ -84,3 +84,5 @@ require (
gopkg.in/natefinch/lumberjack.v2 v2.0.0
gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 // indirect
)

replace github.com/pingcap/parser => github.com/eurekaka/parser v0.0.0-20190717081234-c663361a10d0
4 changes: 2 additions & 2 deletions go.sum
View file
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,8 @@ github.com/dustin/go-humanize v0.0.0-20180421182945-02af3965c54e h1:Fw7ZmgiklsLh
github.com/dustin/go-humanize v0.0.0-20180421182945-02af3965c54e/go.mod h1:HtrtbFcZ19U5GC7JDqmcUSB87Iq5E25KnS6fMYU6eOk=
github.com/eknkc/amber v0.0.0-20171010120322-cdade1c07385 h1:clC1lXBpe2kTj2VHdaIu9ajZQe4kcEY9j0NsnDDBZ3o=
github.com/eknkc/amber v0.0.0-20171010120322-cdade1c07385/go.mod h1:0vRUJqYpeSZifjYj7uP3BG/gKcuzL9xWVV/Y+cK33KM=
github.com/eurekaka/parser v0.0.0-20190717081234-c663361a10d0 h1:34vnlmLZn6QrGRjoUG1J3Cp754kBdd2ZL/4jN+iLvPE=
github.com/eurekaka/parser v0.0.0-20190717081234-c663361a10d0/go.mod h1:IdTOS2WLsXPKKhOm9X8dGbEyWDj8JCYCZqpRZujfsMs=
github.com/fsnotify/fsnotify v1.4.7 h1:IXs+QLmnXW2CcXuY+8Mzv/fWEsPGWxqefPtCP5CnV9I=
github.com/fsnotify/fsnotify v1.4.7/go.mod h1:jwhsz4b93w/PPRr/qN1Yymfu8t87LnFCMoQvtojpjFo=
github.com/ghodss/yaml v1.0.0 h1:wQHKEahhL6wmXdzwWG11gIVCkOv05bNOh+Rxn0yngAk=
Expand Down Expand Up @@ -109,8 +111,6 @@ github.com/pingcap/kvproto v0.0.0-20190429124202-32a5ba2af0f7 h1:+wEqJTc74Jvoxen
github.com/pingcap/kvproto v0.0.0-20190429124202-32a5ba2af0f7/go.mod h1:0gwbe1F2iBIjuQ9AH0DbQhL+Dpr5GofU8fgYyXk+ykk=
github.com/pingcap/log v0.0.0-20190307075452-bd41d9273596 h1:t2OQTpPJnrPDGlvA+3FwJptMTt6MEPdzK1Wt99oaefQ=
github.com/pingcap/log v0.0.0-20190307075452-bd41d9273596/go.mod h1:WpHUKhNZ18v116SvGrmjkA9CBhYmuUTKL+p8JC9ANEw=
github.com/pingcap/parser v0.0.0-20190715020526-98d0068c4ae8 h1:A+hiTQAqlr/LsbEozq14WcSWzgpwSlt4OYoeAedzhzM=
github.com/pingcap/parser v0.0.0-20190715020526-98d0068c4ae8/go.mod h1:1FNvfp9+J0wvc4kl8eGNh7Rqrxveg15jJoWo/a0uHwA=
github.com/pingcap/pd v2.1.12+incompatible h1:6N3LBxx2aSZqT+IWEG730EDNDttP7dXO8J6yvBh+HXw=
github.com/pingcap/pd v2.1.12+incompatible/go.mod h1:nD3+EoYes4+aNNODO99ES59V83MZSI+dFbhyr667a0E=
github.com/pingcap/tidb-tools v2.1.3-0.20190116051332-34c808eef588+incompatible h1:e9Gi/LP9181HT3gBfSOeSBA+5JfemuE4aEAhqNgoE4k=
Expand Down
4 changes: 2 additions & 2 deletions planner/core/expression_rewriter.go
View file
Original file line number Diff line number Diff line change
Expand Up @@ -599,7 +599,7 @@ func (er *expressionRewriter) handleExistSubquery(v *ast.ExistsSubqueryExpr) (as
}
np = er.popExistsSubPlan(np)
if len(np.extractCorrelatedCols()) > 0 {
er.p, er.err = er.b.buildSemiApply(er.p, np, nil, er.asScalar, false)
er.p, er.err = er.b.buildSemiApply(er.p, np, nil, er.asScalar, v.Not)
if er.err != nil || !er.asScalar {
return v, true
}
Expand All @@ -615,7 +615,7 @@ func (er *expressionRewriter) handleExistSubquery(v *ast.ExistsSubqueryExpr) (as
er.err = errors.Trace(err)
return v, true
}
if len(rows) > 0 {
if (len(rows) > 0 && !v.Not) || (len(rows) == 0 && v.Not) {
er.ctxStack = append(er.ctxStack, expression.One.Clone())
} else {
er.ctxStack = append(er.ctxStack, expression.Zero.Clone())
Expand Down

0 comments on commit 6626b49

Please sign in to comment.