Execution plan for queries using minus/union/intersect

Let's say I have select a, b, c from t1 (union/minus/intersect) select a, b , c from t2

Is there any way to avoid the sort unique step (for each table) that takes place before the execution of the set operation?

A unique index maybe or by having the tables sorted by the columns used in the select statement?

I don't have a database around to test it, that's why I'm asking.

Thanks for your help.

-- 
Regards
GBA

--
http://www.freelists.org/webpage/oracle-l