February 2019
Beginner to intermediate
544 pages
14h 36m
English
Spark dataframes have some limitations such as type-safety and therefore the dataset APIs were introduced in version 1.6. The dataset is nothing but an advanced version of a dataframe and is a collection of strongly typed JVM objects. These are represented in tabular format using an encoder. The RDD features such as immutability, fault tolerance, lazy evaluation, and so on are also available with a dataset. The encoder is responsible for the serialization and deserialization of data. It helps in translating data from a JVM object into Spark internal binary format and vice versa. Let's look into a few features of a dataset:
Read now
Unlock full access