Luc-

>This is a complex matter...
>
>I'm working with a legacy software in a Genomic company where they created a database per disease. Each database contain a patient and parent table. When we try to analyse the data in a kind of data warehouse, I need to union the patient and the parent for all disease.
>
>I usually never use view so I try to use the View wizard and was unable to accomplish the task.
>
>So somebody told me to do them programmaticly, It was a good answer but I still hit the limit of 9 union and I need at least 30 unions.
>
>I decided to the drop the view and work with cursor. It won't be dynamic but it will work.

If it really is a data warehouse function, then have you considered creating a denormalized database appropriate for datawarehousing? Or do you still have to support data updates?