[GitHub] incubator-spark pull request: Adding RDD unique self cross product

[rxin]

Github user rxin commented on the pull request:

    https://github.com/apache/incubator-spark/pull/587#issuecomment-34916300
  
    Thanks for submitting this. Just curious, what is the advantage of this 
over rdd.cartesian(rdd), i.e. just use cartesian to join itself?