alamb commented on PR #9445: URL: https://github.com/apache/arrow-rs/pull/9445#issuecomment-6450409016
> I haven’t measured downstream performance, so I can’t claim the impact is negligible. I think preserving per-batch metadata is useful enough to justify this trade-off, but I understand the concern given how widely `RecordBatch` is used. I think I agree with your assesment and I also think it isn't likely to cause problems. That being said there are a LOT of users of RecordBatch. I am imagining someone who hits some memory issue (maybe their query no longer runs when it used to after upgrade) coming to this PR and reading the comments. How do we make sure they feel we have done enough diligence that we tried to avoid causing them problems -- 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]
