Scala:Applied Machine Learning

Chapter 11. Spark SQL and DataFrames

In the previous chapter, we learned how to build a simple distributed application using Spark. The data that we used took the form of a set of e-mails stored as text files.

We learned that Spark was built around the concept of resilient distributed datasets (RDDs). We explored several types of RDDs: simple RDDs of strings, key-value RDDs, and RDDs of doubles. In the case of key-value RDDs and RDDs of doubles, Spark added functionality beyond that of the simple RDDs through implicit conversions. There is one important type of RDD that we have not explored yet: DataFrames (previously called SchemaRDD). DataFrames allow the manipulation of objects significantly more complex than those we have explored to date. ...