rdblue commented on a change in pull request #24768: [SPARK-27919][SQL] Add v2
session catalog
URL: https://github.com/apache/spark/pull/24768#discussion_r299685364
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/UpdateAttributeNullability.scala
##########
@@ -37,7 +37,7 @@ object UpdateAttributeNullability extends Rule[LogicalPlan] {
case p if !p.resolved => p
// Skip leaf node, as it has no child and no need to update nullability.
case p: LeafNode => p
- case p: LogicalPlan =>
+ case p: LogicalPlan if p.childrenResolved =>
Review comment:
If a child is not resolved, then the next line will fail because the
children do not have expression IDs. If the children are never resolved, then
the query will fail the resolution check. If the children are eventually
resolved, then this simply runs later, when the next line will not fail.
This was a preexisting bug that was uncovered by this PR and is caused by
not throwing an analysis exception immediately in `ResolveRelations`. There
have been several updates like this to fix places that didn't check child
resolution.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
