JunRuiLee opened a new pull request, #10966: URL: https://github.com/apache/paimon/pull/10966
### Purpose Ray MERGE INTO counts newly written file rows to report `num_inserted`. With dedicated vector storage, the same logical rows appear in both regular data files and vector files. Only Blob files were excluded, so inserting 5 rows reported 10 even though the committed table contained 5 rows. Exclude `DataFileMeta.is_vector_file()` alongside Blob files when counting inserted rows. Add eight regression cases using real writer messages, commits, and readback, covering inline and dedicated Parquet vectors, with and without Blob columns, across single and multiple write messages. Only distributed message collection is mocked; writing, committing, and reading use the Python implementation. ### Tests - Before the fix: all four dedicated-vector cases reproduced `num_inserted=10` instead of 5; the four inline-vector cases passed. - After the fix: all eight regression cases passed. - Related Parquet vector write and MERGE projection/condition tests: 46 passed, 14 skipped for optional dependencies or version requirements. - Flake8 for both changed files, `git diff --check`, and `mvn -N -o apache-rat:check` passed. Validated against Apache master `dd975e85a8` using Python 3.11, PyArrow 26.0.0, and Ray 2.55.0. -- 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]
