lintingbin opened a new pull request, #4178: URL: https://github.com/apache/amoro/pull/4178
## Why are the changes needed? Close #4122 The `partitions` field in `AbstractOptimizingEvaluator.PendingInput` is annotated with `@JsonIgnore`, which causes partition information to be silently dropped during JSON serialization/deserialization through `TableRuntimeStore` (stored as CLOB in the database). When `PendingInput` is read back from the database, the `partitions` map is always empty. This causes the partition filter in `IcebergTableUtil.createOptimizingPlanner()` to fall back to `Expressions.alwaysTrue()`, meaning all partitions are scanned instead of only the ones that need optimizing. ### Fix Add a serializable `partitionPaths` field (`Set<String>`) that stores partition info as `specId:partitionPath` strings. On serialization, build paths from the in-memory partitions map via `PartitionSpec.partitionToPath()`; on deserialization, rebuild the partitions map from paths using `DataFiles.data()`. ## How was this patch tested? - `TestIcebergTableUtil`: 4 tests passed - `TestOptimizingEvaluator`: 20 tests passed - `TestOptimizingPlanner`: 20 tests passed -- 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]
