Which program is pickling and why? That seems like a very strange thing to do in this situation. The only time to pickle something is if you need to represent a non-string object on disk, but Python ...
I was trying to use Hadoop's streaming pattern to use python code on a largish data set. However, uploading my data to the cluster actually takes approximately forever (I've not yet actually succeeded ...
Results that may be inaccessible to you are currently showing.
Hide inaccessible results