rdblue commented on issue #5591: URL: https://github.com/apache/iceberg/issues/5591#issuecomment-1222524262
This isn't a bug. Metadata and data updates are intended to be separate, although I can see why there are cases where you'd assume that they are not. If you update the schema and commit data in a single job, then it isn't unreasonable to assume the schema change would be rolled back. But if I concurrently add a column while someone else commits, then a rollback should be independent. Expectations can go both ways. While expectations differ, Iceberg never rolls back to a previous schema because that operation is unsafe. For example, if someone deletes a required column and then tries to roll that back, there may have been data written without that column. You can recover the column, but you need to make it optional (or in the future, set a read default). -- 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]
