[GitHub] [arrow-datafusion] tustvold edited a comment on issue #1572: Consolidate the N-way merging code and `SortPreservingMergeStream` (which has quite good tests of what is often quite tricky code, and it will be performance critical)

[GitBox <gi...@apache.org>]

tustvold edited a comment on issue #1572:
URL: https://github.com/apache/arrow-datafusion/issues/1572#issuecomment-1013683820


   > I propose using heap-sort for N-way merge
   > what's your opinion on the merging algorithm to choose?
   
   SGTM :+1:. This was in fact mentioned on the original PR that added the operator [here](https://github.com/apache/arrow-datafusion/pull/379#discussion_r637948151) and led to the filing of #416.
   
   I would describe the current  `SortPreservingMergeStream` somewhat as an MVP, at the time I did not have a need to support a large number of partitions and so went with simple and stupid. I'm aware of some optimisation efforts since then, but unless I've missed something, nothing major. It would be really cool for this component to get some love, I suspect there is substantial potential for performance improvement both in algorithm and implementation :+1: 


-- 
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-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org