[SPARK-1784] Add a new partitioner to allow specifying # of keys per partition

[syedhashmi]

This change adds a new partitioner which allows users
to specify # of keys per partition.

[AmplabJenkins]

Can one of the admins verify this patch?

[srowen]

@rxin Oh interestingly we were just having a side conversation about this.

Is it going to be an issue that the partitioner does not consistently assign the same key to the same partition? I didn't think this would work. The assignments entirely local to the instance of the partitioner. Does this not also maybe yield different result in a multi-threaded context, even with the same input?

[pwendell]

This was merged by mistake by @rxin (I just spoke with him). I am going to revert this. Unfortunately it also broke the master build.

[pwendell]

@rxin merged this in lieu of #729 - so I'm going to merge that one now.

[rxin]

Thanks for doing this.

On Tuesday, May 13, 2014, Patrick Wendell [email protected] wrote:

@rxin https://github.com/rxin merged this in lieu of #729https://github.com/apache/spark/pull/729\- so I'm going to merge that one now.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/721#issuecomment-43046527
.