szehon-ho opened a new pull request, #15743: URL: https://github.com/apache/iceberg/pull/15743
## Summary Follow-up to #15632 (**Core: Fix rewrite_position_delete_files failure with array/map columns**), which only changed the **Spark 4.1** modules. This PR applies the same fix and tests to **Spark 3.4**, **Spark 3.5**, and **Spark 4.0**: - **`PositionDeletesRowReader`**: pass all non-constant field IDs from `expectedSchema().idToName()` into `ExpressionUtil.extractByIdInclusive` (not only primitive columns), so array/map predicates are preserved when reading position delete files. - **Tests**: `testRewritePositionDeletesWithArrayColumns` / `testRewritePositionDeletesWithMapColumns` in `TestRewritePositionDeleteFilesProcedure`, and `testArrayColumnFilter` in `TestPositionDeletesTable`, mirrored from the Spark 4.1 versions. ## Spark versions | Module paths | |--------------| | `spark/v3.4/spark`, `spark/v3.4/spark-extensions` | | `spark/v3.5/spark`, `spark/v3.5/spark-extensions` | | `spark/v4.0/spark`, `spark/v4.0/spark-extensions` | (No changes to Spark 4.1; it already contains #15632.) ## Testing CI should run the Spark 3.4 / 3.5 / 4.0 test tasks for the touched classes. Locally, e.g.: ``` ./gradlew :iceberg-spark:iceberg-spark-3.5_2.13:test --tests "org.apache.iceberg.spark.source.TestPositionDeletesTable.testArrayColumnFilter" ``` -- 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]
