mr-brobot commented on issue #5550: URL: https://github.com/apache/arrow-rs/issues/5550#issuecomment-2025336466
As @tustvold mentioned, this appears to be caused by widening `Int8` and `Int16` columns to `Int32` when writing, then later checking the bloom filter with the unwidened type, resulting in different hashes. I think a general solution would be to apply the same [Arrow-to-Parquet type casting that occurs on write](https://github.com/apache/arrow-rs/blob/ff8611942978c77025286f4051ba20550bef22d4/parquet/src/arrow/arrow_writer/mod.rs#L739-L886) prior to checking the bloom filter. -- 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]
