Spark - RelationalGroupedDataset vs. KeyvalueGroupedDataset? When should I use each of them?

Asked Jan 30 '18 at 05:29

Active Jan 31 '18 at 08:23

Viewed 743 times

When grouping a Dataset in Spark, there are two methods: groupBy and groupByKey[K].

groupBy returns RelationalGroupedDataset, while groupByKey[K] returns KeyvalueGroupedDataset.

What are the differences between them?

Under what circumstances should I choose one over another?

How come my question is a duplicate of those questions about "Dataset vs DataFrame"? I don't get it. It is obviously totally different things! My question is very specific not generic.

edited Jan 31 '18 at 08:23

asked Jan 30 '18 at 05:29

CyberPlayerOne

2,700
3
25
47

8

The title is a duplicate but the actual question is not. Its a shame this was marked as such. – ozzieisaacs Mar 27 '18 at 14:14
1

This question should be re-opened, the answers mentioned don't cover the subject at all – saza Mar 19 '19 at 04:17
This is a valid question and should be re-opened. – reggoodwin May 23 '19 at 08:58

Spark - RelationalGroupedDataset vs. KeyvalueGroupedDataset? When should I use each of them?

0 Answers0