Github user viirya commented on a diff in the pull request: https://github.com/apache/spark/pull/15389#discussion_r82929167 --- Diff: python/pyspark/rdd.py --- @@ -2029,7 +2028,15 @@ def coalesce(self, numPartitions, shuffle=False): >>> sc.parallelize([1, 2, 3, 4, 5], 3).coalesce(1).glom().collect() [[1, 2, 3, 4, 5]] """ - jrdd = self._jrdd.coalesce(numPartitions, shuffle) + if shuffle: + # In Scala's repartition code, we will distribute elements evenly across output + # partitions. However, the RDD from Python is serialized as a single binary data, + # so the distribution fails and produces highly skewed partitions. We need to + # convert it to a RDD of java object before repartitioning. + data_java_rdd = self._to_java_object_rdd().coalesce(numPartitions, shuffle) --- End diff -- @davies Thank you! I do a simple benchmark as above with decreasing the batch size, I don't see an improvement in running time. I.e., import time num_partitions = 20000 a = sc.parallelize(range(int(1e6)), 2) start = time.time() l = a.repartition(num_partitions).glom().map(len).collect() end = time.time() print(end - start) Before: 419.447577953 _to_java_object_rdd(): 421.916361094 decreasing the batch size: 423.712255955 Maybe it depends how is expensive actually converting to java object case by case. Is it generally faster than _to_java_object_rdd()? I would open a followup for this change.
--- 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