jordepic opened a new pull request, #8877: URL: https://github.com/apache/paimon/pull/8877
### Purpose Fixes #8876. The vortex reader classified any projected field named `_SEQUENCE_NUMBER` or `_ROW_ID` as a row-tracking field to synthesize, mapping its projection index to -1. `_SEQUENCE_NUMBER` doubles as the primary-key file format's physical sequence column, so every vortex read of a primary-key table crashed with `ArrayIndexOutOfBoundsException` at the first key-value deserialization. Reserve synthesis for fields the file does not store physically: when the data schema carries the field (by id or name), map it to its physical column like any other. Verified beyond the unit test with an external end-to-end setup: primary-key tables in vortex format under deletion-vector mode, written by paimon-rust and lookup-compacted by this repo's writer, round-trip cleanly with this fix (they crash without it). ### Tests `VortexRecordsReaderTest#testPhysicalSequenceColumnIsReadNotSynthesized`; the two existing schema-handling tests updated for the added data-schema parameter. ### API and Format No changes. ### Documentation No changes. -- 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]
