zhang-arvin opened a new pull request, #18009: URL: https://github.com/apache/iceberg/pull/18009
Fixes #18004. A copy-on-write UPDATE with a WHERE subquery is rewritten by Spark as a UNION of two branches that share the same scan and can invoke the runtime file filter concurrently. SparkCopyOnWriteScan held filteredLocations in a plain field, so a read-then-write race could reset the task list using a stale (single-file) location set — rewriting rows that should not have been touched and duplicating them. This makes filteredLocations an AtomicReference and applies a compare-and-set gate so concurrent filter() calls cannot clobber each other. -- 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]
