Alfred,
why don't you simply create another table with the same structure and a
primary key and then simply do

INSERT INTO NewTable
SELECT * FROM OldTable

That should prevent duplicate values from being assigned (of course, you do
not know which of the duplicate records are inserted).

HTH,
Set

At 12:28 07.05.2003 +0200, you wrote:

>I have a table with no unique primary index. I want to add a unique primary
>key on 2 fields in this table.
>The problem is, the older(garbage data) contains duplicates on these 2
>fields.