andygrove opened a new pull request, #4046:
URL: https://github.com/apache/datafusion-comet/pull/4046
## Which issue does this PR close?
Closes #1839.
## Rationale for this change
With recent Comet DPP work, one of the three DPP tests originally tracked by
#1839 now passes everywhere. The other two reveal a genuine AQE-on Comet DPP
exchange/broadcast reuse bug, now tracked by #4045.
## What changes are included in this PR?
In all three Spark version diffs (3.4.3, 3.5.8, 4.0.1):
- Unignore `avoid reordering broadcast join keys to match input hash
partitioning`.
- Retag `SPARK-32509: ... canonicalization and exchange reuse` and
`SPARK-34637: DPP side broadcast query stage is created firstly` from #1839 →
#4045.
## How are these changes tested?
Verified on Spark 4.0.1 and 3.5.8 with `ENABLE_COMET=true
ENABLE_COMET_ONHEAP=true COMET_PARQUET_SCAN_IMPL=auto` against
`DynamicPartitionPruning{V1,V2,V2Filter}Suite{AEOn,AEOff}`: the unignored test
passes in all six suite variants; the retagged tests still fail with the
patterns described in #4045.
--
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]
