Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/3417#discussion_r20761256
--- Diff: python/pyspark/broadcast.py ---
@@ -49,44 +38,86 @@ def _from_id(bid):
class Broadcast(object):
"""
- A broadcast variable created with
- L{SparkContext.broadcast()<pyspark.context.SparkContext.broadcast>}.
+ A broadcast variable created with L{SparkContext.broadcast()}.
Access its value through C{.value}.
+
+ Examples:
+
+ >>> from pyspark.context import SparkContext
+ >>> sc = SparkContext('local', 'test')
+ >>> b = sc.broadcast([1, 2, 3, 4, 5])
+ >>> b.value
+ [1, 2, 3, 4, 5]
+ >>> sc.parallelize([0, 0]).flatMap(lambda x: b.value).collect()
+ [1, 2, 3, 4, 5, 1, 2, 3, 4, 5]
+ >>> b.unpersist()
+
+ >>> large_broadcast = sc.broadcast(range(10000))
"""
- def __init__(self, bid, value, java_broadcast=None,
- pickle_registry=None, path=None):
+ def __init__(self, sc=None, value=None, pickle_registry=None,
path=None):
"""
- Should not be called directly by users -- use
- L{SparkContext.broadcast()<pyspark.context.SparkContext.broadcast>}
+ Should not be called directly by users -- use
L{SparkContext.broadcast()}
instead.
"""
- self.bid = bid
- if path is None:
- self._value = value
- self._jbroadcast = java_broadcast
- self._pickle_registry = pickle_registry
- self.path = path
+ if sc is not None:
+ f = NamedTemporaryFile(delete=False, dir=sc._temp_dir)
+ self._path = self.dump(value, f)
+ self._jbroadcast =
sc._jvm.PythonRDD.readBroadcastFromFile(sc._jsc, self._path)
+ self._pickle_registry = pickle_registry
+ else:
+ self._jbroadcast = None
+ self._path = path
+
+ def dump(self, value, f):
+ if isinstance(value, basestring):
+ if isinstance(value, unicode):
+ f.write('U')
+ value = value.encode('utf8')
+ else:
+ f.write('S')
+ f.write(value)
+ else:
+ f.write('P')
+ cPickle.dump(value, f, 2)
+ f.close()
+ return f.name
+
+ def load(self, path):
+ with open(path, 'rb', 1 << 20) as f:
+ flag = f.read(1)
+ data = f.read()
+ if flag == 'P':
+ gc.disable()
--- End diff --
Maybe add a comment explaining why you call `gc.disable` here?
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
