>Thanks. I do have over 500,000 records. The clean up will have to be automated.
< shameless plug >
You may wish to consider getting a copy of my Duplicate Record Checker. Right now, it's just on my web page, but in a couple of weeks, I'll upload it to the UT.
< /shameless plug >
>What did you mean by the group columns need to be unique?
When one says that a set of records is duplicated, that means that some predefined combination of fields acts as a key to the table, and only one record may hold any given set of values. For example, if I say that my primary key in a table is SSN and Birthdate, then only one record will have SSN = 333-44-6666 and Birthdate = 12/04/1974, and if more than one record matches that key, then those records are considered to be duplicates. If one wishes to have a primary (unique) key in a table, then the key may be as small as one column or as large as all the columns in the table.
Who knows, it may be that your table may have 500,000 records, but there are only 4 duplicated keys in it, with 5 records for each key. :)