bert-beyondloops commented on PR #26261:
URL: https://github.com/apache/datafusion/pull/26261#issuecomment-6853505729
Hi @andygrove and @comphead , could you take a look when you have time?
You both worked on the SortMergeJoin perf changes (#26277, #26192), so I'd
value your read on this.
This PR does two things:
1. **Derives SortMergeJoin sort options from the input ordering.** The
join keys are an unordered set of equalities, so the planner now reorders them,
together with their sort options, to match the ordering both inputs
already declare. This avoids adding a `SortExec` on inputs that are
already sorted that way (see `sort_merge_join_sort_options.slt`).
2. **Allows SortMergeJoin with `prefer_hash_join=false` for a single
target partition.** Before, it was only chosen when `target_partitions > 1` and
`repartition_joins` was on.
--
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]
