zhuqi-lucas commented on PR #186: URL: https://github.com/apache/datafusion-site/pull/186#issuecomment-4998603495
Thanks so much @alamb — went through everything and your edits read really well. The restructure (moving buffering / decoder-loop into appendices, tightening the intro, expanding Related Work) makes the whole post much clearer, and the phrasing is consistently better than what I had. Pushed one tiny follow-up commit ([1666952](https://github.com/apache/datafusion-site/pull/186/commits/1666952)) with three small technical fixes on top of your work: 1. **Row-Level Early Stopping paragraph**: attributed the "skip projection column fetch" behavior to arrow-rs's `RowFilter` (where it actually lives) and clarified that #22450's contribution is keeping the threshold up-to-date at row-group boundaries so this fires more often. 2. **"In DataFusion 54"** → "Over the last few DataFusion releases" — #22450 merged 2026-06-22, after `54.0.0` was tagged, so it will ship in a later release; safer to not commit to a version. 3. **Decision-tree figure alt-text**: `try_pushdown_sort` → `PushdownSort` to match the struct name in the surrounding prose. On your two suggestions: - **Raw numbers for sort_tpch**: agreed, I'll add per-query timing table like the sort_pushdown section has. Will do in a follow-up. - **Appendix structure**: happy with your move, keeps the main narrative flowing. +1 on aiming for **Monday July 20** publish; I'll ping on Discord over the weekend once the per-query table lands so anyone else who wants a last read can weigh in. -- 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]
