pyspark.SparkContext.broadcast#

SparkContext.broadcast(value)[source]#

Broadcast a read-only variable to the cluster, returning a Broadcast object for reading it in distributed functions. The variable will be sent to each cluster only once.

Added in version 0.7.0.

Parameters:

valueT: value to broadcast to the Spark nodes

Returns:

Broadcast: Broadcast object, a read-only variable cached on each machine

Examples

>>> mapping = {1: 10001, 2: 10002}
>>> bc = sc.broadcast(mapping)

>>> rdd = sc.range(5)
>>> rdd2 = rdd.map(lambda i: bc.value[i] if i in bc.value else -1)
>>> rdd2.collect()
[-1, 10001, 10002, -1, -1]

>>> bc.destroy()