parthchandra commented on issue #7040: URL: https://github.com/apache/arrow-rs/issues/7040#issuecomment-2626072993
Yeah. As @tustvold said we might be looking at the undefined behavior part of the spec. @tustvold the parquet implementation I was referring to is the one pointed to by the Apache Parquet project - https://github.com/apache/parquet-java/ @etseidl the pyarrow output matches my expectation. Afaik, pyarrow uses [arrow-cpp's parquet](https://arrow.apache.org/docs/python/parquet.html) These two implementations are probably the most widely used implementations out there. > One could also argue it's a bug in the java implementation...when writing an 8 bit unsigned, you'd expect values to be masked to the proper number of bits to avoid this issue And therein lies the rub :) . -- 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]
