We have read and understood well that RDDs are an immutable, distributed collection of object values used as a unit of abstraction in Spark Framework. There are two ways RDDs can be created:
- Loading external dataset
- Distributing a list/set/collection of objects in their driver program
Now let's create some simple programs to create and use RDDs:
The preceding screenshot captures quick steps to create an RDD on Spark Shell. Here are the specific commands and further transformational outputs for this:
Scala> val inputfile = sc.textFile("input.txt")
The preceding command reads the file called input.txt from the specified absolute location and a new RDD is created under the name inputfile. In the preceding snippet we have not specified the entire path, thus the framework would assume that the file exists under the current location...
