Hi,
Sean Davis schrieb:
> Probably the simplest way to do this is to load the data into a temporary
> table without the unique constraint then use SQL to insert a "clean" version
> into the new table. Alternatively, you could use savepoints. If an insert
> fails, just rollback to that savepoint. If it succeeds, commit that
> savepoint.
>
But isn't that more overhead than using the trigger? And while inserting
into the clean table, I would have to test if the value is existing or
not. I don't see the advantage of this compared to the script which was
posted a few weeks ago.
Regards,
Verena
