ulysses-you commented on code in PR #36177:
URL: https://github.com/apache/spark/pull/36177#discussion_r850050641
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/joins.scala:
##########
@@ -184,6 +202,23 @@ object EliminateOuterJoin extends Rule[LogicalPlan] with
PredicateHelper {
case a @ Aggregate(_, _, p @ Project(_, Join(_, right, RightOuter, _, _)))
if a.groupOnly && p.references.subsetOf(right.outputSet) =>
a.copy(child = p.copy(child = right))
+
+ case a @ Aggregate(_, aggExprs, Join(left, _, LeftOuter, _, _))
+ if a.references.subsetOf(left.outputSet) &&
+ allDuplicateAgnostic(aggExprs) =>
+ a.copy(child = left)
+ case a @ Aggregate(_, aggExprs, Join(_, right, RightOuter, _, _))
+ if a.references.subsetOf(right.outputSet) &&
+ allDuplicateAgnostic(aggExprs) =>
+ a.copy(child = right)
+ case a @ Aggregate(_, aggExprs, p @ Project(_, Join(left, _, LeftOuter, _,
_)))
Review Comment:
We would pull out nondeterministic to a new project, so some thing like this
?
```sql
SELECT rand(), max(c2) FROM (
SELECT t1.c1, t1.c2 FROM t1 LEFT JOIN t2 ON t1.c1 = t2.c1
)
GROUP BY rand()
```
How about adding a condition `a.groupingExpressions.forall(_.deterministic)`
?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
