taoran92 opened a new pull request, #10652: URL: https://github.com/apache/paimon/pull/10652
### Purpose Fixes #10651. Selecting `__paimon_bucket`, or using it only in a filter, fails when a chain table reads snapshot and delta data through a `ChainSplit`. Resolve bucket metadata from the underlying `DataSplit` or `ChainSplit` while preserving the original fallback and query-authorization wrappers for reading. Derive the chain bucket ID from existing bucket paths without changing split serialization. Partition metadata remains restricted to the original `DataSplit`, preserving the authorization boundary for raw partition values. ### Tests - `ChainSplitTest` and `ChainTableUtilsTest`: 27 tests passed. - `SplitUtilsQueryAuthTest`: 4 tests passed. - `SparkChainTableMetadataTest`: both parameterized cases passed, covering multiple buckets, projections, bucket-only filters, snapshot-only reads, snapshot/delta merges, and query authorization. - Both Spark regression cases fail with the original split-type exception when using the parent commit's production classes. The changed production and test sources were compiled into a temporary directory against existing local build dependencies and tested with JUnit Platform and ScalaTest (Spark 3.5.8, Scala 2.12.18, Java 11). A full Maven build and Spark 4 tests were not run. The following validation passed without `fast-build`: ```sh mvn -o -pl paimon-core,paimon-spark/paimon-spark-common,paimon-spark/paimon-spark-ut -Pspark3 -DskipTests validate ``` -- 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]
