2010YOUY01 commented on PR #15380: URL: https://github.com/apache/datafusion/pull/15380#issuecomment-2811725054
I really like this idea, I find this approach will not conflict with several future optimizations in my mind: - Use row format for sorting and reuse converted `Row`s for SPM, this only requires changing the current `vec<SortColumn>` to a single `Row`, and won't introduce a structure change. - Eagerly `Row`s conversion for more robust memory accounting https://github.com/apache/datafusion/issues/14748 Additionally, `sort_in_place_threshold_bytes` is relatively easy to tune: it can be set as a typical cache size. If the number of sorted runs is still too large after combing, a cascaded merge can still be applied on top of that. I think it's a good idea to proceed. -- 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: github-unsubscr...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org
