GitHub user singularperturbation opened a pull request:
https://github.com/apache/spark/pull/15670
[SPARK-18161] [Python] Allow use of pickle to serialize > 4 GB objects
where available
## What changes were proposed in this pull request?
Since pickle protocol 4 (available in Python 3.4+) allows for serialization
of > 4 GB errors, use pickle protocol 4 when available in the Broadcast and
PickleSerializer classes.
## How was this patch tested?
This was tested through pyspark shell and serializing a large numpy array
with / without the changes.
Author: Sloane Simmons <sloanes dot k at gmail dot com>
This contribution is my original work and I license the work to the project
under the project's open source license.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/singularperturbation/spark pickle_protocol_4
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/15670.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #15670
----
commit 571d01b25a6f6f84e5dfe0551f2b08fcfeee132e
Author: singularperturbation <[email protected]>
Date: 2016-10-28T08:11:30Z
Use protocol 4 if available for PickleSerializer
commit cd5f8c76bdd89091c180ad54dd1d26d9a0a2cc3f
Author: singularperturbation <[email protected]>
Date: 2016-10-28T08:11:54Z
Use protocol from serializers in Broadcast as well
----
---
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]
