Spark is designed to overcome the limitations of Hadoop and MapReduce paradigm in

general. List at least two types of algorithms where Spark outperforms Hadoop. $[2$ PTs$]$

List the three methods that could create a Spark RDD$.[1\mathrm{.}5$ PTs$,0\mathrm{.}5$ PT each$]$

Spark's operations are divided into two classes $[3$ PTs$,1$ PT each$]$:

a$)$ List the name of each class.

b$)$ Give an example of each class.

c$)$ What is the main difference between these two classes?

The following snippet is the $($Word Count example$)$ written to be executed by Spark!

val. textFile $=$ sc$.$textFile$("$hdfs:$//...")$

val counts $=$ textFile,flathap $($line $=>$ line,split$(""$i$))$

$.$ map $($word $=>($word$,1))$

$.$reduceByKey$(++)$

counts. saveAsTextFile $("$hdfs:$//...")$

Answer the following questions based on the code above $[3$ PTs$]$:

a$)$ What is the method used to create the RDD called "textFile"?

b$)$ List all the transformations used above.

c$)$ What is the action method used above?