Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/18652#discussion_r127895248
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
---
@@ -1912,6 +1913,26 @@ class Analyzer(
nondeterToAttr.get(e).map(_.toAttribute).getOrElse(e)
}.copy(child = newChild)
+ case j: Join if j.condition.isDefined &&
!j.condition.get.deterministic =>
+ j match {
+ // We can push down non-deterministic joining keys.
+ // We can't push down non-deterministic conditions.
+ case ExtractEquiJoinKeys(_, leftKeys, rightKeys, conditions, _,
_)
--- End diff --
Joining keys can only be equi-join. It is exactly the use case discussed in
the dev mailling list. It's actually useful for the use cases.
A general non-deterministic join condition pushdown doesn't make a lot of
sense. The kind of predicates like `rand(1) > 0 && rand(11) < 0` can be a
serious concern. The join results can be different before and after pushdown.
---
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.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
