Github user mateiz commented on a diff in the pull request: https://github.com/apache/spark/pull/156#discussion_r11266448 --- Diff: core/src/main/scala/org/apache/spark/rdd/RDD.scala --- @@ -664,6 +664,24 @@ abstract class RDD[T: ClassTag]( } /** + * Return a Stream that contains all of the elements in this RDD. + * + * In case of iterating it consumes memory as the biggest partition in cluster. + */ + def toLocallyIterable: Stream[T] = { + def collectPartition(p: Int): Array[T] = { + sc.runJob(this, (iter: Iterator[T]) => iter.toArray, Seq(p), allowLocal = false).head + } + var buffer = Stream.empty[T] + for (p <- 0 until this.partitions.length) { + buffer = buffer append { + collectPartition(p).toStream --- End diff -- Actually it would probably be cleaner to write this code like this: ``` (0 until partitions.length).iterator.flatMap(i => collectPartition(i)) ``` Since we're not returning a Stream anymore, but rather an iterator. You could also just make `collectPartition` take a Partition object and then do `partitions.iterator.flatMap(...)`.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---