Skip to content

fix: enable ParquetFilterSuite tests for native_datafusion#3698

Closed
andygrove wants to merge 2 commits intoapache:mainfrom
andygrove:fix/enable-native-datafusion-parquet-filter-tests
Closed

fix: enable ParquetFilterSuite tests for native_datafusion#3698
andygrove wants to merge 2 commits intoapache:mainfrom
andygrove:fix/enable-native-datafusion-parquet-filter-tests

Conversation

@andygrove
Copy link
Member

Which issue does this PR close?

Closes #3320.

Rationale for this change

Two ParquetFilterSuite tests were disabled for native_datafusion mode due to a CometNativeExec should not be executed directly without a serialized plan crash. This crash was fixed by #3511 which added a dedicated doExecuteColumnar() to CometNativeScanExec, so these tests no longer need to be skipped.

What changes are included in this PR?

Regenerated dev/diffs/3.5.8.diff to remove IgnoreCometNativeDataFusion tags from two ParquetFilterSuite tests:

  • "SPARK-31026: Parquet predicate pushdown for fields having dots in the names"
  • "Filters should be pushed down for Parquet readers at row group level"

Both tests have their stripSparkFilter assertions already guarded by if (!isCometEnabled || isCometScanOnly), so they pass safely when Comet native execution is active.

The other two tests from the issue were already resolved:

  • "filter pushdown - StringPredicate" is tagged with IgnoreCometNativeScan("cannot be pushed down") for a separate valid reason
  • "SPARK-41017" already has Comet scan type pattern matching in the existing diff

How are these changes tested?

By running the Spark SQL tests with Comet enabled in native_datafusion mode, which exercises the re-enabled ParquetFilterSuite tests.

Remove IgnoreCometNativeDataFusion tags from two ParquetFilterSuite
tests that were disabled due to a CometNativeExec crash when executing
without a serialized plan. The crash was fixed by PR apache#3511 which added
a dedicated doExecuteColumnar() to CometNativeScanExec. The test
assertions are already guarded by isCometEnabled/isCometScanOnly checks.
@andygrove andygrove marked this pull request as ready for review March 14, 2026 15:47
@andygrove andygrove requested a review from mbutrovich March 14, 2026 15:56
Copy link
Contributor

@mbutrovich mbutrovich left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @andygrove! I love to see more tests running with native_datafusion!

@andygrove andygrove marked this pull request as draft March 16, 2026 20:44
@andygrove
Copy link
Member Author

moving to draft to avoid accidental merge - I need to rebase and merge each of the native_datafusion PRs individually to avoid corrupting the diff

@andygrove andygrove closed this Mar 17, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[native_datafusion] [Spark SQL Tests] CometNativeExec crash — executed without serialized plan

2 participants