Github user ravipesala commented on the issue:
https://github.com/apache/carbondata/pull/2971
@QiangCai @jackylk Adding a `RANGE_COLUMN` at each load level does not
create an issue? If user selects different range column for each load how you
are going to compact when you support it in future?
what is the background of giving the range_column in load level instead of
create table level?---
