CODEWITHSUNDEEP
scalaapache-sparkapache-spark-sqlrdd
sarthak
sarthak

Reputation: 794

Case statements in Spark

I'm writing a Spark code where I need to convert an RDD of type (String,(String,String)) to ((String,String),String).

I have the following input text file:

Language,Language-code,TotalViewsInThatLang
English,en,10965376,"Main_Page",2938355
Russian,ru,1925718,"%D0%97%D0%B0%D0%B3%D0%BB,915495
Spanish,es,1010810,"Wikipedia:Portada",13603

I have created an RDD as follows:

val line = sc.textFile(inputFile)
val nrdd = line.map(x=>(x.split(",")(0),(x.split(",")(1),x.split(",")(2))))
nrdd: org.apache.spark.rdd.RDD[(String, (String, String))] = MapPartitionsRDD[2] at map at <console>:26

From this I want to use casefunction to create RDD of type ((String,String),String).

How can I do this with case statements in map?
EDIT
I am getting the following error when I am trying to use case function:

scala> val frdd = nrdd.map( {case(x,(y,z))=>((x,y),z))}) 
<console>:1: error: ';' expected but ')' found.
val frdd = nrdd.map({case(x,(y,z))=>((x,y),z))})
                                             ^

Upvotes: 2

Views: 4447

Answers (2)

Shankar
Shankar

Reputation: 8957

Since your RDD is Pair RDD, You could use swap from Keyed RDD.

Sample Code:

val keyRDD = sc.parallelize(List((("a1", "b1"), "c1"), (("a2", "b2"), "c2")), 2)
val swappedRDD = keyRDD.map(_.swap)
swappedRDD.foreach(x => println(x))

Upvotes: 0

slouc
slouc

Reputation: 9698

Unless I misunderstood your question, you want this:

val list: List[((String, String), String)] = List((("a1", "b1"), "c1"), (("a2", "b2"), "c2"))
val res = list.map { case ((a, b), c) => (a, (b, c)) }

println(res) // List((a1,(b1,c1)), (a2,(b2,c2)))

Upvotes: 3

Related Questions