Hi Jackie,
I assume you're using Oracle. If so, this should do the trick:
It uses the rowid pseudo-column to resolve the duplicates. If you're not using Oracle you can achieve the same result using description instead of rowid, but it's obviously not as efficient to join on a char/varchar.
One small caveat is that I don't have an RDBMS here to
test it, but I'm pretty confident it's correct. It should at least give you an idea of how to proceed.
You can probably also achieve the same effect in Oracle using a dynamic view (sub-query in the FROM clause) but it makes my brain hurt trying to visualise it without anywhere to test it!
Hope that helps.
Jules