f46e02fcdb
For shuffle-based operators, such as rdd.groupBy() or rdd.sortByKey(), PySpark will always assume that the default parallelism to use for the reduce side is ctx.defaultParallelism, which is a constant typically determined by the number of cores in cluster. In contrast, Spark's Partitioner#defaultPartitioner will use the same number of reduce partitions as map partitions unless the defaultParallelism config is explicitly set. This tends to be a better default in order to avoid OOMs, and should also be the behavior of PySpark. JIRA: https://issues.apache.org/jira/browse/SPARK-2203 Author: Aaron Davidson <aaron@databricks.com> Closes #1138 from aarondav/pyfix and squashes the following commits: 1bd5751 [Aaron Davidson] SPARK-2203: PySpark defaults to use same num reduce partitions as map partitions |
||
---|---|---|
.. | ||
mllib | ||
__init__.py | ||
accumulators.py | ||
broadcast.py | ||
cloudpickle.py | ||
conf.py | ||
context.py | ||
daemon.py | ||
files.py | ||
java_gateway.py | ||
join.py | ||
rdd.py | ||
rddsampler.py | ||
resultiterable.py | ||
serializers.py | ||
shell.py | ||
sql.py | ||
statcounter.py | ||
storagelevel.py | ||
tests.py | ||
worker.py |