On 13/02/2026 22:15, Matheus Alcantara wrote:
> Yeah, I agree. I think that we can be less restrictive. The behavior can
> be to allow reference objects from schemas different than source and in
> case some error happen it stop the process. This behavior should be
> documented.
The more I think about it, the more it seems like the safest approach.
As I mentioned upthread, the user might otherwise end up with pretty
much useless tables and the pleasure of having to drop them manually :)
> In case of e.g table partitions which the parent is in different schema
> the overlap error will happen and the the process will stop. For another
> cases like types in different schema it will work normally.
+1
Best, Jim
