Github user squito commented on the issue: https://github.com/apache/spark/pull/19848 i dunno what the requirements are -- I was hoping you would know the hadoop committer semantics better than me! I suppose a uuid is really the only get something globally unique, as you could even have multiple independent spark contexts. I have seen a committer creating a temp directory based on the ID, so you could end up with a collision with them both writing to the same dir. anyway, I'm willling to set this aside as a rare case, the fix here is still a huge improvement.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org