limadog9 commented on PR #25818: URL: https://github.com/apache/datafusion/pull/25818#issuecomment-5913933959
Thanks @alamb for the review and for starting CI! I've pushed d8c9d051a to address your feedback: - Avoided constructing errors for expected unsupported bound types and timestamp overflow. These now discard unusable bounds without allocating an error message, while preserving unrelated statistics. - Extracted and documented the distinct-count preservation check, reusing `CastExpr::check_bigger_cast`. The tests now include `Float64` to `Float32` narrowing and integer widening alongside timestamp conversions. - Clarified what the metadata tests cover beyond the SQL regression: `COPY` cannot produce explicit distinct-count statistics or mixed timestamp-bound exactness flags. The failure cases also check that both bounds are cleared while independent counts are retained. All 269 Parquet tests, `parquet_sorted_statistics.slt`, formatting, and focused Clippy with warnings denied passed locally. I've updated the PR description with the validation details and limitations; the broader workspace checks and benchmark comparison remain incomplete locally. This is ready for another look whenever you have a chance. The new workflow runs are awaiting maintainer approval again. Thanks for your time and guidance! -- 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]
