Skip to content

[CALCITE-4693] Query with Lateral Join should converted to Left Join … #4474

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
NobiGo wants to merge 3 commits into
base: main
Choose a base branch
from
Draft

[CALCITE-4693] Query with Lateral Join should converted to Left Join … #4474

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
04c414d
[CALCITE-4693] Query with Lateral Join should converted to Left Join …
NobiGo Jul 22, 2025
b7278f3
[CALCITE-4693] Code Review
NobiGo Jul 23, 2025
f55e538
[CALCITE-4693] Add Unit Test can't Pass
NobiGo Jul 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 23 additions & 0 deletions core/src/main/java/org/apache/calcite/sql/validate/SqlValidatorImpl.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1418,6 +1418,29 @@ private void handleOffsetFetch(@Nullable SqlNode offset, @Nullable SqlNode fetch
if (node instanceof SqlMerge) {
validatingSqlMerge = true;
}
if (node instanceof SqlJoin) {
// select * from "scott".dept,
// lateral (select count(cast(deptno as integer)) from "scott".emp
// where emp.deptno = dept.deptno);
// ==》
// select * from "scott".dept left join
// lateral (select count(cast(deptno as integer)) from "scott".emp
// where emp.deptno = dept.deptno) on true;
SqlJoin sqlJoin = (SqlJoin) node;
if (sqlJoin.getJoinType() == JoinType.COMMA && sqlJoin.getRight() instanceof SqlCall) {
Copy link
Contributor

@mihaibudiu mihaibudiu Jul 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you explain where the discussion in the JIRA is reflected in this code?
I think the consensus there was that a left join is not always right

Copy link
Contributor

@julianhyde julianhyde Jul 24, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree, @mihaibudiu. The jira cases discusses how converting an inner lateral join to a left join is WRONG. Do not merge this PR.

Copy link
Contributor Author

@NobiGo NobiGo Jul 25, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I will recheck this. This repair method does have some issues. I will add more test cases to verify it.

SqlCall sqlCall = (SqlCall) sqlJoin.getRight();
if (sqlCall.getOperator() == SqlStdOperatorTable.LATERAL
&& sqlCall.operand(0) instanceof SqlSelect
&& isAggregate(sqlCall.operand(0))) {
return performUnconditionalRewrites(
new SqlJoin(sqlJoin.getParserPosition(),
sqlJoin.getLeft(), sqlJoin.isNaturalNode(),
JoinType.LEFT.symbol(node.getParserPosition()),
sqlJoin.getRight(), JoinConditionType.ON.symbol(node.getParserPosition()),
SqlLiteral.createBoolean(true, node.getParserPosition())), underFrom);
}
}
}
SqlCall call = (SqlCall) node;
final SqlKind kind = call.getKind();
final List<SqlNode> operands = call.getOperandList();
Expand Down
94 changes: 94 additions & 0 deletions core/src/test/resources/sql/lateral.iq
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -246,4 +246,98 @@ where job = 'MANAGER';

!ok

# [CALCITE-4693] Query with Lateral Join should converted to Left Join When sub-query is aggregate query

select * from "scott".dept, lateral (select count(cast(deptno as integer)) from "scott".emp where emp.deptno = dept.deptno);
+--------+------------+----------+--------+
| DEPTNO | DNAME | LOC | EXPR$0 |
+--------+------------+----------+--------+
| 10 | ACCOUNTING | NEW YORK | 3 |
| 20 | RESEARCH | DALLAS | 5 |
| 30 | SALES | CHICAGO | 6 |
| 40 | OPERATIONS | BOSTON | 0 |
+--------+------------+----------+--------+
(4 rows)

!ok
EnumerableCalc(expr#0..4=[{inputs}], expr#5=[IS NULL($t4)], expr#6=[0:BIGINT], expr#7=[CASE($t5, $t6, $t4)], proj#0..2=[{exprs}], EXPR$0=[$t7])
EnumerableHashJoin(condition=[=($0, $3)], joinType=[left])
EnumerableTableScan(table=[[scott, DEPT]])
EnumerableCalc(expr#0..2=[{inputs}], expr#3=[IS NOT NULL($t2)], expr#4=[0], expr#5=[CASE($t3, $t2, $t4)], DEPTNO=[$t0], EXPR$0=[$t5])
EnumerableNestedLoopJoin(condition=[IS NOT DISTINCT FROM($0, $1)], joinType=[left])
EnumerableCalc(expr#0..2=[{inputs}], DEPTNO=[$t0])
EnumerableTableScan(table=[[scott, DEPT]])
EnumerableAggregate(group=[{0}], EXPR$0=[COUNT($1)])
EnumerableCalc(expr#0..7=[{inputs}], expr#8=[CAST($t7):INTEGER], expr#9=[IS NOT NULL($t7)], DEPTNO=[$t7], $f0=[$t8], $condition=[$t9])
EnumerableTableScan(table=[[scott, EMP]])
!plan

select * from "scott".dept, lateral (select count(cast(deptno as integer)) from "scott".emp where emp.deptno = dept.deptno + 10);
+--------+------------+----------+--------+
| DEPTNO | DNAME | LOC | EXPR$0 |
+--------+------------+----------+--------+
| 10 | ACCOUNTING | NEW YORK | 5 |
| 20 | RESEARCH | DALLAS | 6 |
| 30 | SALES | CHICAGO | 0 |
| 40 | OPERATIONS | BOSTON | 0 |
+--------+------------+----------+--------+
(4 rows)

!ok
EnumerableCalc(expr#0..5=[{inputs}], expr#6=[IS NULL($t5)], expr#7=[0:BIGINT], expr#8=[CASE($t6, $t7, $t5)], proj#0..2=[{exprs}], EXPR$0=[$t8])
EnumerableHashJoin(condition=[=($3, $4)], joinType=[left])
EnumerableCalc(expr#0..2=[{inputs}], expr#3=[10], expr#4=[+($t0, $t3)], proj#0..2=[{exprs}], $f3=[$t4])
EnumerableTableScan(table=[[scott, DEPT]])
EnumerableCalc(expr#0..2=[{inputs}], expr#3=[IS NOT NULL($t2)], expr#4=[0], expr#5=[CASE($t3, $t2, $t4)], DEPTNO0=[$t0], EXPR$0=[$t5])
EnumerableNestedLoopJoin(condition=[IS NOT DISTINCT FROM($0, $1)], joinType=[left])
EnumerableAggregate(group=[{0}])
EnumerableCalc(expr#0..2=[{inputs}], expr#3=[10], expr#4=[+($t0, $t3)], $f3=[$t4])
EnumerableTableScan(table=[[scott, DEPT]])
EnumerableAggregate(group=[{0}], EXPR$0=[COUNT($1)])
EnumerableCalc(expr#0..7=[{inputs}], expr#8=[CAST($t7):INTEGER], expr#9=[IS NOT NULL($t7)], DEPTNO0=[$t8], $f0=[$t8], $condition=[$t9])
EnumerableTableScan(table=[[scott, EMP]])
!plan

select * from "scott".dept, lateral (select sum(cast(deptno as integer)) from "scott".emp where emp.deptno = dept.deptno);
+--------+------------+----------+--------+
| DEPTNO | DNAME | LOC | EXPR$0 |
+--------+------------+----------+--------+
| 10 | ACCOUNTING | NEW YORK | 30 |
| 20 | RESEARCH | DALLAS | 100 |
| 30 | SALES | CHICAGO | 180 |
| 40 | OPERATIONS | BOSTON | |
+--------+------------+----------+--------+
(4 rows)

!ok
EnumerableCalc(expr#0..4=[{inputs}], proj#0..2=[{exprs}], EXPR$0=[$t4])
EnumerableMergeJoin(condition=[=($0, $3)], joinType=[left])
EnumerableTableScan(table=[[scott, DEPT]])
EnumerableSort(sort0=[$0], dir0=[ASC])
EnumerableAggregate(group=[{0}], EXPR$0=[SUM($1)])
EnumerableCalc(expr#0..7=[{inputs}], expr#8=[CAST($t7):INTEGER], expr#9=[IS NOT NULL($t7)], DEPTNO=[$t7], $f0=[$t8], $condition=[$t9])
EnumerableTableScan(table=[[scott, EMP]])
!plan

select * from "scott".dept, lateral (select sum(cast(deptno as integer)) from "scott".emp where emp.deptno = dept.deptno + 10);
+--------+------------+----------+--------+
| DEPTNO | DNAME | LOC | EXPR$0 |
+--------+------------+----------+--------+
| 10 | ACCOUNTING | NEW YORK | 100 |
| 20 | RESEARCH | DALLAS | 180 |
| 30 | SALES | CHICAGO | |
| 40 | OPERATIONS | BOSTON | |
+--------+------------+----------+--------+
(4 rows)

!ok
EnumerableCalc(expr#0..5=[{inputs}], proj#0..2=[{exprs}], EXPR$0=[$t5])
EnumerableHashJoin(condition=[=($3, $4)], joinType=[left])
EnumerableCalc(expr#0..2=[{inputs}], expr#3=[10], expr#4=[+($t0, $t3)], proj#0..2=[{exprs}], $f3=[$t4])
EnumerableTableScan(table=[[scott, DEPT]])
EnumerableAggregate(group=[{0}], EXPR$0=[SUM($1)])
EnumerableCalc(expr#0..7=[{inputs}], expr#8=[CAST($t7):INTEGER], expr#9=[IS NOT NULL($t7)], DEPTNO0=[$t8], $f0=[$t8], $condition=[$t9])
EnumerableTableScan(table=[[scott, EMP]])
!plan

# End lateral.iq
Loading