Github user sujith71955 commented on the issue: https://github.com/apache/spark/pull/22758 > I think the cost of get the stats from `HadoopFileSystem` may be quite high. Then we shall depend on HiveStats always to get the statistics, which is happening now also but partially. and i think this PR solving that problem, But what i told is based on cloudFans expectation ![image](https://user-images.githubusercontent.com/12999161/47195764-f3874700-d37a-11e8-9b93-e3c1cb228c54.png)
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org