Datasets, DataFrames and RDDs¶

Many may have been asking yourself why they should be using Datasets rather than the foundation of all Spark - RDDs using case classes.

This document collects advantages of Dataset vs RDD[CaseClass] to answer the question Dan has asked on twitter:

"In #Spark, what is the advantage of a DataSet over an RDD[CaseClass]?"

Saving to or Writing from Data Sources¶

With Dataset API, loading data from a data source or saving it to one is as simple as using SparkSession.read or Dataset.write methods, appropriately.

Accessing Fields / Columns¶

You select columns in a datasets without worrying about the positions of the columns.

In RDD, you have to do an additional hop over a case class and access fields by name.