Github user kumarvishal09 commented on the issue:
https://github.com/apache/carbondata/pull/2252
@xuchuanyin Its better to add one encoder type to store long string. In
current code reader does not know which type of data it is reading/storing and
chunk store object is created based on encoder (fixed/variable) type. In your
PR most of the classes you have added one boolean to check its Long string.
It's not required, you can add one encoder type(Text) and instead of handling
everything in same
class(UnsafevariableLengthChunkStore/SafevariableLengthStore) add one more
implementation for handling Long String.
---
