>Michel said that it produced an error, but I gather that you are saying this particular error doesn't trigger an Error event. If so, that's a COUNT command bug I didn't know about, but guess an aggregate SELECT command could generally be used as a workaround.

But, then again, you will have to set a new corporate rule for all developers to not use Count. And, whenever you do it, there will be always someone who will do it.

When trying to solve that issue this morning, we realize that the implication could be extremly severe as the code continue as is. So, if an error occurs, the rest of the code could populate bad data for example.