May 2017
Intermediate to advanced
294 pages
7h 33m
English
Datasets were added to Spark 1.6 to provide strong typing to DataFrames. In Spark 2.0, Dataset and the DataFrame API were merged to provide one single abstraction. At the bare level, a DataFrame can be considered a Dataset of Row objects. This will be enhanced in various places, as you will see in subsequent chapters (for example, in Machine Learning a Dataset of (label, BestVector)).
Datasets provide a rich set of transformation APIs, such as RDDs. It is recommended that you use Datasets as much as possible.
Read now
Unlock full access