YannByron opened a new pull request, #11155: URL: https://github.com/apache/paimon/pull/11155
### Purpose The Python deferred BLOB reader bypasses FileIO range coalescing when `blob_parallelism` is 1, including the default. It reads each descriptor separately, even when descriptors point to adjacent or overlapping ranges in the same file. Use the existing `read_blobs_concurrent` path for one worker as well. This lets serial reads merge ranges and reuse a stream while preserving result order, duplicate descriptors, nulls, table-scoped FileIO, and configured merge limits. For the regression fixture, the default path changes four individual payload reads into one merged range using one stream. This change is limited to the Python deferred scalar BLOB reader; the Rust reader and the configured parallelism are unchanged. ### Tests - `PYPAIMON_TEST_EXECUTION=python python -m pytest -q pypaimon/tests/deferred_blob_resolve_test.py pypaimon/tests/blob_test.py`: **205 passed** (Python 3.12.14, PyArrow 19.0.1). - The three new parameterized cases fail against the unmodified upstream implementation and pass with the change. They cover default merging, a zero-gap limit, a maximum merged-span limit, duplicate/out-of-order descriptors, nulls, and closing the reused stream. - Flake8 with the repository configuration passed for all three changed files. -- 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]
