Github user squito commented on a diff in the pull request: https://github.com/apache/spark/pull/11105#discussion_r55446400 --- Diff: core/src/main/scala/org/apache/spark/scheduler/ResultTask.scala --- @@ -66,7 +67,10 @@ private[spark] class ResultTask[T, U]( _executorDeserializeTime = System.currentTimeMillis() - deserializeStartTime metrics = Some(context.taskMetrics) - func(context, rdd.iterator(partition, context)) + val itr = rdd.iterator(partition, context) + val result = func(context, itr) + val computedFullPartition = itr.isEmpty || rdd.storageLevel != StorageLevel.NONE --- End diff -- hmm, doesn't this need to know whether we've computed the full partition for *all* the rdds involved in this task, not just the last one? Eg., what if an earlier rdd is cached, but the final one is not. Or what if there is a coalesce, so the coalesced partition is not computed fully, but the partitions of the original rdds are computed fully? :( kinda worried this is a critical issue ...
--- 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. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org