select distinct over specific columns

A query in a system I maintain returns

QID AID DATA
1   2   x
1   2   y
5   6   t

As per a new requirement, I do not want the (QID, AID)=(1,2) pair to be repeated. We also dont care what value is 开发者_运维问答selected from "data" column. either x or y will do.

What I have done is to enclose the original query like this

SELECT * FROM (<original query text>) Results group by QID,AID

Is there a better way to go about this? The original query uses multiple joins and unions and what not, So I would prefer not to touch it unless its absolutely necesary

---

If you don't care which DATA will be selected, GROUP BY is nice, though using ungrouped and unaggregated columns in SELECT clause of a GROUP BY statement is MySQL specific and not portable.