akshaytayal opened a new pull request, #13580: URL: https://github.com/apache/gluten/pull/13580
## What changes are proposed in this pull request? Spark 4.3 `UnionExec` (SPARK-57881) reports a merged `KeyedPartitioning` whose partitions are the children's partitions concatenated, so `numPartitions` is their sum. `ColumnarUnionExec` passed it to `SQLPartitioningAwareUnionRDD`, which interleaves same-index partitions, and failed with `ArrayIndexOutOfBoundsException`. - Pass the union's `Partitioning` to the `SparkContextUtils.createPartitioningAwareUnionRDD` shim. - On Spark 4.3, concatenate the children with `sparkContext.union` for `KeyedPartitioning`, as Spark's `UnionExec` does; other partitionings still use `SQLPartitioningAwareUnionRDD`. - Spark 3.4–4.2 behaviour is unchanged. Re-enabled Spark 4.3 tests: - `GlutenGroupBasedUpdateTableSuite`: 12 - `GlutenGroupBasedRowLevelOperationCatalystRuntimeFilterSuite`: 1 - `GlutenKeyGroupedPartitioningSuite`: 4 Depends on #13520 Related issue: #13517 ## How was this patch tested? Existing Spark 4.3 UTs re-enabled above. ## Was this patch authored or co-authored using generative AI tooling? Generated-by: GitHub Copilot CLI (Claude Opus 5.5) -- 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]
