>There are other ways to do this. Getting duplicate key values from a table is a simple SQL SELECT statement, and another statement afterwards can be used to delete all but one...

What statement do you use to delete all but one?
I am sometimes a bit afraid that using SQL against a corrupted table would generate wrong results, but if you make sure no index is used, I guess this would work too and I might use this next time.

I don't really see the advantage of this method over the unique index method, expect that maybe you would completely want to get rid of unique indexes. I agree that they are in fact quite useless, and I have used other ways to repair tables without using unique indexes. I just use it in this case because I think it is easy. I wouldn't mind if there were no unique indexes though and I agree that they shouldn't be used normally.

Auke