I have a question.
I'm running a spark (scala) job via Oozie workflow and its giving me a slightly different result, whereas running the same spark code via spark-shell gives me the expected output.
While there is no change in the code, what makes the output different? Have any of you faced a situation like this in your project? Appreciating your comments.
P.S: I'm using spark 2.x. I'm not facing any error though while running as a Oozie job. Just the output difference is my concern.