thswlsqls opened a new pull request, #10379: URL: https://github.com/apache/paimon/pull/10379
### Purpose fix #10378 - Java table reads lose matching rows when `withReadType` prunes a nested field that a filter references (`s.b = 7` with `s ROW<a>`), without `executeFilter()`. - #9858 drops filters on unprojected columns but checks only top-level names; after #9423 enabled nested pushdown, `s.b` passes and Parquet reads it as all-null. - `ParquetReaderFactory` now also requires the whole nested path in the read type. AND keeps projected conjuncts; an OR touching a pruned path is dropped. - Only the Java `ReadBuilder` path is affected. Flink/Spark SQL keep filter fields in the read type; `executeFilter()` is already safe. ### Tests - Added `AppendOnlySimpleTableTest#testParquetFilterOnUnprojectedNestedField`; fails on master. - Added `ParquetReadWriteTest#testReadWithUnprojectedNestedFilter` (AND, OR); both fail on master. - `ParquetReadWriteTest` 54/54 and `AppendOnlySimpleTableTest` 114/114 pass; checkstyle and spotless pass on `paimon-format` and `paimon-core`. -- 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]
