andygrove opened a new issue, #4513: URL: https://github.com/apache/datafusion-comet/issues/4513
## What is the problem the feature request solves? Comet can route eligible Spark `ScalaUDF` expressions through its Arrow-direct codegen dispatcher, keeping the project, exchange, and sort operators around a UDF on the native Comet path instead of falling back to Spark. This is gated behind `spark.comet.exec.scalaUDF.codegen.enabled`, which currently defaults to `false` and is documented as experimental. The feature now has broad type coverage (scalars, complex types with arbitrary nesting, higher-order functions) and end-to-end correctness, fuzz, and Iceberg test coverage. It should be enabled by default so users benefit without opt-in. ## Describe the potential solution - Flip the default of `spark.comet.exec.scalaUDF.codegen.enabled` to `true`. - Remove the "experimental" status from the config doc and user guides. - Document how to disable the dispatcher for users who want the previous behavior. ## Additional context Users can still set `spark.comet.exec.scalaUDF.codegen.enabled=false` to restore the prior fall-back-to-Spark behavior. -- 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]
