Github user xuchuanyin commented on the issue:
https://github.com/apache/carbondata/pull/2685
1. I think there may be some bugs in index datamap currently to support
'different sort columns per load'
2. What about different dictionary_include per load?---
