unikdahal opened a new issue, #6611: URL: https://github.com/apache/datafusion-comet/issues/6611
## Problem `MergeRowsExec` repeatedly evaluates predicates and filters/materializes batches as rows move through multiple MERGE clauses. This becomes increasingly expensive for wider inputs and multi-clause MERGEs. ## Proposed solution - track remaining rows using selection vectors - evaluate later predicates only on unclaimed rows - gather only predicate-referenced columns - materialize each fired action once - preserve zero-copy slices where possible Add native and Spark-level benchmarks to validate both correctness and performance. Related: #5318, #5122 -- 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]
