Github user HyukjinKwon commented on the issue:
https://github.com/apache/spark/pull/21537
I am sorry if I misunderstood. This was discussed sufficiently in the
linked PRs. These are internal APIs and does not introduce behaviour changes by
AnalysisBarrier which was a good enough idea to try because it brought many
benefits to reduce reanalyzing plans - this resolved many JIRA on the other
hand, and I don't think AnalysisBarrier was particularly a disaster. Sometimes
we should take a step forward and see if that causes an actual problem or not
too. That was a good enough try.
When something should be reverted, there should usually be specific
concerns; otherwise, I wouldn't revert this for vague concerns for now.
This is an internal API that we can freely change later as well. Also,
@viirya made a lot of efforts here and he's pretty active. I don't think
dragging someone into this is a good idea for now.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
