Scala Code not executing

问题:

I am trying to execute the following scala code on Spark but due to some reason the function selective is not getting called

var lines = sc.textFile(inputPath+fileName,1)
val lines2 =lines.map(l=>selective(
func,List(2,3),List(1,26),l,";",0           
))
lines2.toArray().foreach(l=>out.write(l))
.......

The selective function is defined as follows

def selective(f: (String,Boolean) => String , phoneFields: Seq[Int], codeFields: Seq[Int], in: String, delimiter:String, iMode:Int /* 0 for enc, 1 for dec */) :String =
        in.split(delimiter,-1).zipWithIndex
    .map { 

      case (str, ix) 
    if( phoneFields.contains(ix)||codeFields.contains(ix)) => 
        var output=f(str,codeFields.contains(ix))
        var sTemp=str+":"+output+"\n"
        if((iMode==0)&&codeFields.contains(ix)&&(str.compareTo("")!=0)  )
            CodeDictString+=sTemp
        else if(str.compareTo("")!=0)
            PhoneDictString+=sTemp

        output
      case other => other._1
    }.mkString(";").+("\n")

The println statement is not executing. Furthermore the function is not returning any thing. sc is the spark context object

回答1:

This function does not compile. The syntax

{
  some statement
  case ... => ...

is not valid. Case statements can only appear like this:

{
   case ... => ...
       ...
   case ... => ...
       ...
}

Since you obviously got something to compile, I bet there's a case statement before that println, and that case statement is not being selected.

回答2:

Are you running this in local mode or on a cluster? The function passed to lines.map is evaluated by the Spark workers, so that println will appear in the worker's stdout logs if you're running on a cluster (these logs are viewable via Spark's web UI).