Github user jongyoul commented on the issue:
https://github.com/apache/zeppelin/pull/3100
@zjffdu Yes, it has a very long story. I, personally, don't want to keep
both of them, but didn't find any solution to fit both requirements. `-Pr` is
pure R interpreter which supported - I didn't test it - spark inside it. AFAIK,
no one tests if SparkR works for now with `-Pr`.---
