I ran into the same exception using the Data Frame and changing the cluster configuration didn't do any help. I tried some of the suggestions from the above but those didn't either.

The only way I could get around this use was to create a temporary view from the Data Frame and do a select with only a limited number of results on it. After this I was able to use the entire temporary view without any issues. If I don't do a LIMIT on the number of results I ran into the same issue again even the with the view. Hope this helps someone.