Github user cloud-fan commented on the issue: https://github.com/apache/spark/pull/13909 sorry for the delay. Yea looks like we can't reuse the byte array of unsafe data in expressions, which may get cached expectedly and leads to wrong result. I'm a little concerned about the hacks in `BufferHolder` and the array writer. The code is so coupled with unsafe row writer and we have to hack it so that we can write unsafe array directly. What if we actually write an unsafe row with a single array field and return the array column? Then we don't need the hacks, but waste some bits for the row format overhead, which seems acceptable.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org