pyspark.streaming.DStream.join#

DStream.join(other, numPartitions=None)[source]#

Return a new DStream by applying ‘join’ between RDDs of this DStream and other DStream.

Hash partitioning is used to generate the RDDs with numPartitions partitions.