>If I understand your question try...
>
>SELECT COUNT(*) as keycount, keyfield FROM table1 GROUP BY upper(keyfield) ORDER BY upper(keyfield) HAVING keycount >1
>
>Also, if you have a large table, make sure that you have and index on upper(keyfield)
>Hope this helps //:^)
>

I do not know about SQL Server, but you can not create an upper(keyfield) index in Oracle. So optimization is no good there. Better to convert data to upper, lower or proper at the time of data entry. Then there is no need to be concerned about haw many different variations a user can enter text data.