andygrove commented on issue #2065:
URL: 
https://github.com/apache/datafusion-ballista/issues/2065#issuecomment-4994192735

   Root-caused and fixed in #2066. The promotion replaces the inner 
`SortMergeJoinExec` with a `HashJoinExec(CollectLeft)`, which does not preserve 
input order, leaving the parent SMJ to merge unsorted input.
   
   The fix re-establishes any hard input ordering a rewritten child no longer 
carries, and adds the suggested validation: every `SortMergeJoinExec` input 
must satisfy the join's required ordering, so this surfaces as an error rather 
than wrong results. The check reports no false positives across the full SF1 
TPC-DS gate under both `prefer_hash_join=true` and `false`.
   
   It also fixes q4, so with #2066 the gate goes 86 → 88.


-- 
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]

Reply via email to