>Thank you for the reply. I kind of got the expected results using cursor set processing but your approach is much nicer. I will try to get it to work using your idea.
>
>Thanks again,
>Daniel

I was just modifying my reply to get farther - I don't want to use union, I would use two CTEs instead, etc.

Anyway, hopefully you get the idea and will post your final result here - I will switch back to the Mobile app noboby wants to help me with here today...