Github user xuchuanyin commented on the issue:
https://github.com/apache/carbondata/pull/2324
@ravipesala @jackylk
I'm not OK with the default size and fpp for bloomfilter. In general
usecases, it will result in many false samples.
In my test, a 96GB customer table with size=320K & fpp=0.00001 will result
in 18 false samples.---
