if I understand what you want basically you should treat this as a coumpund index you may have to do some type conversions but assuming that both fields are character thus, something like SELECT keyfieild,field2, recno() as REC() ; from Table where KEYfield+field2 in (SELECT Keyfield+field2 as CompKey from Table Group BY 1 Having COUTNT(compkey) >1 should work Arnon > >Can someone help me with the SQL to find duplicate records in a table where > a duplicate is considered two out of five matching fields? I thought: >SELECT * ; >FROM table; >WHERE keyfield IN ; >(SELECT keyfield ; >FROM tablel ; >GROUP BY keyfield, field2; >HAVING count(*) > 1) >would do the trick, but it doesn't work. TIA > >Erik > >