JeonDaehong opened a new pull request, #3590: URL: https://github.com/apache/iceberg-rust/pull/3590
## Which issue does this PR close? - Closes #3464. ## What changes are included in this PR? parquet-rs truncates footer statistics to 64 bytes by default, so a position delete file whose data file path is longer than that gets no `file_path` bounds. Readers (`referenced_data_file` in the delete file index, and Java's `ContentFileUtil.referencedDataFile`) then can't tell it covers a single data file and apply it to every data file in the partition. `ParquetWriterBuilder::build` now turns statistics truncation off when the schema has the reserved `file_path` field, like Java's `MetricsConfig.forPositionDelete`. Other files keep the configured truncation. #3537 doesn't cover this, since truncated bounds are still not equal. Opening as a draft for now because of the open PR limit. I'll mark it ready once one of mine is merged. ## Are these changes tested? Yes. `test_position_delete_writer_keeps_long_file_path_bounds` writes 40, 70 and 200 byte paths (and 200 with an explicit truncate length of 16) and checks both bounds equal the path. It fails without the fix. `test_statistics_truncation_kept_off_for_position_deletes` checks a data file schema keeps its truncate length. Clippy, fmt, docs, msrv and the full nextest run are clean. ## AI Disclosure I used Claude Code to write the change and run the tests on Linux. I reviewed it before opening the PR. -- 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]
