Hi Sergey, thanks for the reply.

>The BINARY_CHECKSUM() function would return the same value if rows are the same (duplicates) and it ignores text, ntext and image columns. You can temporay add a calculated column with BINARY_CHECKSUM() as expression to your table and use it as starting point in finding duplicates. Unfortuantelly, there's as small chance that the different data in a row can produce the same checksum so you'll have to add some other columns to the comparison.
>Keep in mind that it's just an idea and test thoroughly.

Mmmmm... more complicated than I thought. I might just end-up importing the table into a VFP cursor ordered by pk, doing a SCAN/ENDSCAN, deleting each 2nd instance of a record, then doing a tableupdate (after deleting all records in the original SQL table so I don't get a pk update conflict).

Regards,
Alan