Github user kumarvishal09 commented on the issue:
https://github.com/apache/carbondata/pull/2252
@xuchuanyin I think some of the changes are not required ...like method
added for converting to LV format..If u check direct compressor it is already
present ..you can use the same.
In DimensionRawColumnChunk no need to pass any Boolean based on encoder we
can decide which type of dimension store object needs to be created
Changing the existing store chunk implementation is also not required
...add a child class if possible or add complete new implementation for storing
int based LV format...
Please check ColumnPage.compress and ColumnPage.decompress for you
reference(LV)
V1,V2 reader no need to change anything as its old format code and user
will not able to load the data in this format.
Decide based on encoder. while writing add new encoder "TEXT" and while
reading use same encoder for creating DimensionDataChunkStore object
---
