>>We tried it, but it's slower, than the methods described. Less code is not always correspond to the most efficient method.
>
>OK, I agree. But I still think you should mention the alternative. You can mention, of course, that it is slower - and give some examples.
>
>For example, sometimes I want to find orphan records only once, that is, I don't need to run a query every day. That is, once I find the offending records, I will turn on RI. And to run the orphan-query only once, or a few times, I don't really care if my code takes 10 seconds instead of 2. So I think there are uses for both approaches - the INDEXSEEK() approach (more code, and faster), and the subquery approach (less coding, but slower).
>
>Hilmar.

I don't want to run bencmarking tests again, though...

I finally was able to update it, it would appear in the new version soon.