peterolkhov opened a new pull request, #1010: URL: https://github.com/apache/arrow-nanoarrow/pull/1010
Sliced sparse unions currently select child values from earlier rows because their child-run cursor omits the parent offset. For example, slicing `[10, 'twenty', 30, 'forty']` at offset 2 produces `[10, 'twenty']` instead of `[30, 'forty']`. Start sparse child runs at the effective parent offset and read exactly the requested number of type IDs in both union iterators. The exact-length correction also prevents an extra row when a dense union is a child of a sliced struct. Adds regression coverage for sparse/dense slices, custom type codes, nullable children, child offsets, empty slices and nested list/struct/RecordBatch conversion. Tested locally on macOS arm64/Python 3.14/PyArrow 26: 276 Python tests passed (4 skipped), 118 doctests passed, stubtest passed for 8 modules, applicable configured pre-commit hooks passed. Built and installed the changed source wheel with two compilation jobs. Ubuntu/Python-version matrix and sdist/twine CI were not run locally. AI assistance: OpenAI Codex generated the patch and tests. -- 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]
