Github user jiangxb1987 commented on the issue: https://github.com/apache/spark/pull/22112 The changes looks good from my side, it summarizes the current insight we have towards the data correctness issue caused by input order aware operators and inconsistent shuffle output order, also it provides a temporarily workaround of the above issue by failing. I feel we can have this in 2.4 and continue investigation in future releases. Let's listen to @tgravescs @mridulm @markhamstra who have been actively tracking the issue to see whether we can move forward with this PR?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org