I had tried that.  The table is 8 million rows right now and even trying to limit it to a specific unique ID the server basically locks up on the query (the cube goes to using 110+ GB memory and doesn't return the result so I end up having to kill the cube).  It's an annoying issue.