Gabriel39 commented on PR #67784: URL: https://github.com/apache/doris/pull/67784#issuecomment-5679725556
Review follow-up in 78ded3952e: - Renamed the Paimon and Iceberg writer entry points to write_column_to_paimon_arrow and write_column_to_iceberg_arrow throughout the SerDe hierarchy. - Kept versioned mixed-format Iceberg scans on V2 even when the default write format is ORC and retained Parquet ranges arrive later. - Isolated condition-cache entries by the normalized INT96 timezone contract, with a real-file red/green regression for 2048 incorrectly skipped rows. Local validation: 239 tests across 9 BE suites passed, including all 4 new regressions; the BE test executable rebuilt successfully; all 13 affected production objects compiled; clang-format 16 passed for all 86 C++ files changed by this PR. The full diff was self-reviewed before committing. A first broad run encountered a pre-existing temporary-directory permission collision; the complete test selection passed using an isolated temporary directory, without changing or excluding that test. Four review threads are addressed. The VARBINARY SQL compatibility discussion remains open: restoring optional STRING mapping changes the stated catalog-mapping goal, while implementing every missing SQL consumer substantially expands this PR. This commit does not claim to resolve that compatibility concern. -- 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]
