Shekharrajak opened a new pull request, #3075: URL: https://github.com/apache/datafusion-comet/pull/3075
Fixes https://github.com/apache/datafusion-comet/issues/2944 This happens because `RangeExec` (and other non-Comet Spark operators) produce Spark's `OnHeapColumnVector` instead of Arrow arrays that the native writer expects. ## What changes are included in this PR? - Modified `CometNativeWriteExec.doExecuteColumnar()` to detect when the child operator is not a `CometPlan` - Added automatic conversion of Spark columnar batches to Arrow format using `CometArrowConverters.columnarBatchToArrowBatchIter()` - Added support for row-based input by converting rows to Arrow batches using `CometArrowConverters.rowToArrowBatchIter()` ## How are these changes tested? Added two new tests in `CometParquetWriterSuite`: -- 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]
