[SPARK-23207][SQL] Shuffle+Repartition on a DataFrame could lead to incorrect answers#20393

jiangxb1987 wants to merge 2 commits intoapache:masterapache/spark:masterfrom

jiangxb1987:shuffle-repartitionjiangxb1987/spark:shuffle-repartitionCopy head branch name to clipboard