Rich-T-kid commented on issue #12197: URL: https://github.com/apache/arrow-rs/issues/12197#issuecomment-6827996770
Closing this alternative. Stopping at an encoding boundary avoids the immediate spill, but it changes batch sizing and can move the work into downstream coalescing, filtering, and repartitioning. That is not the behavior we want from the Parquet reader. The preferred direction is #12195: keep the requested batch shape and avoid expanding existing dictionary rows by merging unique values and remapping keys across dictionary/plain transitions. -- 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]
