cloud-fan commented on PR #40462: URL: https://github.com/apache/spark/pull/40462#issuecomment-1484415289
The optimizer can remove meanless logical operators as it won't change the query result and it's usually more efficient to remove operators. But for physical APIs like rebalance and repartition, the users explicitly ask to take control of performance tuning and we should trust them. Besides, I don't think this is an always-beneficial optimization, what if the limit is very large? We should let the user make the decision: they can remove the rebalance hint if they believe it makes the perf worse. What do you think? @wangyum @dongjoon-hyun -- 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]
