GitHub user JoshRosen opened a pull request:
https://github.com/apache/spark/pull/11498
[SPARK-13653] Split disk writer into separate object and binary writers
This patch splits `DiskBlockObjectWriter` into two separate classes so that
the object-based and byte-based write methods are cleanly separated. The old
design, which put all of these methods in the same interface, caused some weird
problems in places which only wanted to use the writer's `OutputStream`
methods, since those places would need to pass a `DummySerializerInstance` into
the writer's constructor.
This patch introduces a `DiskBlockWriter` which exposes only the
OutputStream write methods and modifies the `DiskBlockObjectWriter` to wrap
`DiskBlockWriter` and expose only object-based write methods.
I also moved the disk-writing-related classes into a new
`o.a.s.storage.disk` package, since the `o.a.s.storage` package had too many
classes in it.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/JoshRosen/spark disk-writer-cleanup
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/11498.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 #11498
----
----
---
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]
