I encountered a scenario where I am getting incorrect hive query results in program executed using Spark-submit. When I execute the same query using Hue and on command line of Spark-shell using HiveContext I get correct results and it matches, but when I execute the same using the Spark program usinf Spark-Submit it gave incorrect results. The table properties are listed below and the table has around 248M rows. Any particular reason for this to happen? Any workaround or config settings that needs to be done?