yupp, but I think this might be one of the cases where xBase gives you
better or at least more stable execution times across diverse data distributions.
(guessing that no index exists on empty(adressX))

I'd probably code it as xBase first, if tasked with it.

regards

thomas

>Seems pretty straightforward to me Message ID: 1516033
>
>>I believe that is a bit too complicated to do with SQL.
>>
>< snip >