diff options
author | Liang-Chi Hsieh <viirya@gmail.com> | 2016-10-11 11:43:24 -0700 |
---|---|---|
committer | Felix Cheung <felixcheung@apache.org> | 2016-10-11 11:43:24 -0700 |
commit | 07508bd01d16f3331be167ff92770d19c8b1f46a (patch) | |
tree | c9c4e44b1908c44a0f9235d2c60ced1587201c06 /external/flume-assembly/pom.xml | |
parent | 75b9e351413dca0930e8545e6283874db09d8482 (diff) | |
download | spark-07508bd01d16f3331be167ff92770d19c8b1f46a.tar.gz spark-07508bd01d16f3331be167ff92770d19c8b1f46a.tar.bz2 spark-07508bd01d16f3331be167ff92770d19c8b1f46a.zip |
[SPARK-17817][PYSPARK] PySpark RDD Repartitioning Results in Highly Skewed Partition Sizes
## What changes were proposed in this pull request?
Quoted from JIRA description:
Calling repartition on a PySpark RDD to increase the number of partitions results in highly skewed partition sizes, with most having 0 rows. The repartition method should evenly spread out the rows across the partitions, and this behavior is correctly seen on the Scala side.
Please reference the following code for a reproducible example of this issue:
num_partitions = 20000
a = sc.parallelize(range(int(1e6)), 2) # start with 2 even partitions
l = a.repartition(num_partitions).glom().map(len).collect() # get length of each partition
min(l), max(l), sum(l)/len(l), len(l) # skewed!
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. We need to convert the RDD in Python to java object before repartitioning.
## How was this patch tested?
Jenkins tests.
Author: Liang-Chi Hsieh <viirya@gmail.com>
Closes #15389 from viirya/pyspark-rdd-repartition.
Diffstat (limited to 'external/flume-assembly/pom.xml')
0 files changed, 0 insertions, 0 deletions