How do I return Spark RDD partition values without a local iterator?

Question

I'm learning Spark and its parallelism that relates to RDD partition distributions. I have a 4 CPU machine hence I have 4 units of parallelism. To return the members of partition index "0" I couldn't find a way to return this partition without forcing the RDD to use a localIterator.

I'm used to spark being quite terse. Is there a more concise way to filter an RDD by partition? The following two methods work, but it seems clumsy.

scala> val data = 1 to 20
data: scala.collection.immutable.Range.Inclusive = Range(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)

scala> val distData = sc.parallelize(data)
distData: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[75] at parallelize at <console>:26

scala> distData.mapPartitionsWithIndex{
   (index,it) => {
      it.toList.map(x => if (index == 0) (x)).iterator
   }
}.toLocalIterator.toList.filterNot(
   _.isInstanceOf[Unit]
)
res107: List[AnyVal] = List(1, 2, 3, 4, 5)

scala> distData.mapPartitionsWithIndex{
   (index,it) => {
      it.toList.map(x => if (index == 0) (x)).iterator
   }
}.toLocalIterator.toList.filter(
   _ match{
      case x: Unit => false
      case x => true
   }
)
res108: List[AnyVal] = List(1, 2, 3, 4, 5)

Answer 1

distData.mapPartitionsWithIndex{ (index, it) => 
      if (index == 0) it else Array[Int]().iterator
}

You can return an empty iterator and it will work fine.