matthewgapp commented on issue #7636: URL: https://github.com/apache/arrow-datafusion/issues/7636#issuecomment-1737571569
@alamb thanks for the continued review and for the heads up on the removal of that invariant from arrow. We should still ensure that the nullability information is correct at all times, even if arrow doesn't complain. I completely agree that changing the column generated by ROW_NUMBER() should **_not_** be marked as `nullable: true`. To that end, my PR (#7638) specifically introduces `nullable: false` **_when it should be false_** for window functions in the logical plan. The code in main assumes that the column generated by all window functions is `nullable: true` which is wrong. In other words, my PR brings the logical plan in line with the actual schema produced by the execution plan. Let me know if I'm missing something. -- 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]
