Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/2351#discussion_r17992850
--- Diff: python/pyspark/rdd.py ---
@@ -2081,8 +2085,44 @@ def _jrdd(self):
self.ctx.pythonExec,
broadcast_vars,
self.ctx._javaAccumulator)
self._jrdd_val = python_rdd.asJavaRDD()
+
+ if enable_profile:
+ self._id = self._jrdd_val.id()
+ if not self._created_profiles:
+ dump_path = self.ctx._conf.get("spark.python.profile.dump")
+ if dump_path:
+ atexit.register(PipelinedRDD.dump_profile, dump_path)
+ else:
+ atexit.register(PipelinedRDD.show_profile)
+ self._created_profiles.append((self._id, profileStats))
+
return self._jrdd_val
+ @classmethod
+ def show_profile(cls):
--- End diff --
What do you think about introducing a new class to hold the PySpark
profiling methods? We could call it something like `Profiling` and put it in
`profiling.py`.
It seems a little weird to allow `show_profile()` to be called on a
particular RDD instance when the method prints all created profiles.
---
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]
