Github user fjh100456 commented on the issue: https://github.com/apache/spark/pull/19218 Thanks for your review. @gatorsmile In the first question I mean that âparquet.compressionâ can be found in the `table: Tabledesc` (maybe similar with `catalogtable`), and can also be found in `sparkSession.sessionState.conf`(set by user through the command `set parquet.compression=xxx`), which one should take precedence? This issue was originally only related to hive table writing, but after fix the priority, it was found that non-partitioned tables did not take the right precedence, and non-partitioned tables writing will not enter `InsertIntoHiveTable`. `Insertintohadoopfsrelationcommand.scala` is really not a proper place, is there any place that can solve both of partitioned tables and non-partitioned tables?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org