September 2018
Intermediate to advanced
398 pages
9h 43m
English
The core concept of Spark is the RDD. From the user's point of view, for a given type A, RDD[A] looks similar to a standard Scala collection, such as Vector[A]: they are both immutable and share many well-known methods, such as map, reduce, filter, and flatMap.
However, the RDD has some unique characteristics. They are as follows:
Lazy: When you call a transformation function, such as map or filter, nothing happens immediately. The function call is just added to a computation graph that is stored in the RDD class. This computation graph is executed when you subsequently call an action function, such as collect or take.
Distributed: The data in the RDD is split in several partitions that are scattered across ...
Read now
Unlock full access