YjyJeff opened a new issue, #7834: URL: https://github.com/apache/arrow-datafusion/issues/7834
### Describe the bug The [comment](https://github.com/apache/arrow-datafusion/blob/main/datafusion/core/src/physical_optimizer/enforce_distribution.rs#L1047) in `EnforeDistribution` said: > Hashing benefits from partitioning, add a round-robin repartition before it It sounds pretty odd. In a specific example: ``` before: AggregateExec: mode=FinalPartitioned, gby=[ip@0 as ip], aggr=[COUNT(Int64(1))] AggregateExec: mode=Partial, gby=[ip@0 as ip], aggr=[COUNT(Int64(1))] after: AggregateExec: mode=FinalPartitioned, gby=[ip@0 as ip], aggr=[COUNT(Int64(1))] RepartitionExec: partitioning=Hash([ip@0], 2), input_partitions=2 RepartitionExec: partitioning=RoundRobinBatch(2), input_partitions=1 AggregateExec: mode=Partial, gby=[ip@0 as ip], aggr=[COUNT(Int64(1))] ``` Can anyone explain why the `RepartitionExec: partitioning=RoundRobinBatch(2), input_partitions=1` could accelerate the plan? In my view, removing this `RepartitionExec` will be faster 😂 Thanks in advance. ### To Reproduce _No response_ ### Expected behavior _No response_ ### Additional context _No response_ -- 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]
