Ann W. Harrison wrote:

> Radu Sky wrote:
>
>> If you need to remove duplicate rows within unnamed cursor you can try
>> using a new UNIQUE field.
>
> Or use the always present, always unique rdb$db_key.
>

My intention was to use the unique field to generate the -803 exception,
to exclude duplicates.
How can this be achieved using rdb$db_key ?

Regards,

Radu