andygrove opened a new issue, #3067: URL: https://github.com/apache/datafusion-comet/issues/3067
### What is the problem the feature request solves? Native shuffle does not currently support round-robin partitioning, so Comet will fall back to the JVM shuffle instead. It isnt' necessarily a high priority to implement this, but I thought it would be good to have an issue to track this and discuss the design. ### Describe the potential solution When `spark.sql.execution.sortBeforeRepartition` is enabled (which is the default) then we need to compute a murmur3 hash for all columns in the input data and then sort the data before partitioning. This is needed so that the partitioning is deterministic, which is important in the event of a task being retried. ### 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
