HeartSaVioR commented on issue #23782: [SPARK-26875][SS] Add an option on FileStreamSource to include modified files URL: https://github.com/apache/spark/pull/23782#issuecomment-555273404 > File lastest_sales.csv gets deleted (manually or via the new configuration #22952) I don't think it is safe to delete the file manually outside of Spark, as it cannot be known whether the file is processed and added to the commit log. Checking whether the file is processed is not sufficient - it should ensure the file will be never accessed. Assuming the file is deleted/archived from Spark, I'm seeing the possible confusion from user side as well since it's no longer an "overwrite" if the file is processed and cleaned up. So that seems to be a valid use case, though we may need to think how we let Spark differentiate "overwrite" and "add new file which was deleted or moved".
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
