[SPARK-4636] [SQL] Cluster By & Distribute By should follows the Hive behavior#3496

chenghao-intel wants to merge 1 commit intoapache:masterapache/spark:masterfrom

chenghao-intel:repartitionchenghao-intel/spark:repartitionCopy head branch name to clipboard