Hi All I found that impala already implemented Zorder ( https://issues.apache.org/jira/browse/IMPALA-8755).
I used to think supporting zorder needed file format support, but from the impala implementation, it looks like only needing to implement a new RecordComparator which is independent with File format. I know that DRB already supports it in 2018, I believe this feature is useful for improving query performance in certain cases. Would it be possible to port the zorder implementation from DBR to open source spark? Thanks Chang