Github user HeartSaVioR commented on the issue: https://github.com/apache/spark/pull/21721 Unfortunate thing is that continuous mode allows different epochs between partitions, hence query progress just doesn't fit so hard to address SPARK-23887 for now. My 2 cents is continuous mode also should sync epoch in query so all partitions deal with same epoch. When we deal with aggregation it should be essential.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org