Selecting unique count on non-unique index

Hi All, I've spotted a table in our database that has no unique key but is nevertheless updated using a WHERE clause that identifies a uniqueness. That is: UPDATE a_table SET a_field = a_value WHERE a_field1 = a_value1 AND a_field2 = a_value2 AND a_field3 = a_value3 The fields a_field1, a_field2, a_field3 have not been collated to form a unique key ( which I intend to do! ). Having found the total number of records on the table, how do I now determine the number of records based on the uniqueness of the three fields sited above? Cheers, Ian. Tedious 'office speak' part 3: "Let's go scuba in the think tank."