Github user zfong commented on a diff in the pull request:
https://github.com/apache/drill/pull/748#discussion_r101089177
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/planner/physical/NestedLoopJoinPrule.java
---
@@ -49,7 +49,7 @@ protected boolean checkPreconditions(DrillJoinRel join,
RelNode left, RelNode ri
PlannerSettings settings) {
JoinRelType type = join.getJoinType();
- if (! (type == JoinRelType.INNER || type == JoinRelType.LEFT)) {
+ if (!(type == JoinRelType.INNER || (type == JoinRelType.LEFT &&
JoinUtils.hasScalarSubqueryInput(left, right)))) {
--- End diff --
I see. So, Calcite is converting a query with a subquery into an
equivalent left join with a scalar subquery. Do you know if the query returns
the correct result in that case if a nested loop join is used? If not, then it
seems like we should still be returning an error.
Also, how do you distinguish the case where the query has been converted to
this form by Calcite vs a user explicitly doing a left outer join to a scalar
subquery. It seems like in this case, if Drill uses a nested loop join, it
will also return wrong result.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
