T1 <-- T2 <-- T3
Is it possible to define a unique constraint involving one column from T3 and one from T1?

e.g. Customer has multiple Application(s). Application has multiple Ticket(s). The TicketNumber should be unique for every Customer.

Thanks & Regards.

what are the primary keys for those tables?

In the example,
Customer - CustomerId (Natural Key)
Application - Surrogate Key
Ticket - Surrogate Key

presumably you're using an autoincrement or summat similar?

then it will be unique for every customer because it's unique, full stop

Sorry, I will attempt to make it a little more clear.

TicketId is unique because it is autoincrement. However there is an additional attribute TicketNumber which is what I need to make unique within a customer. That is, I need to allow same ticket number across customers (C1-T1 and C2-T1), but need to ensure uniqueness within a customer (C1-T1 is unique). In summary, the CustomerId+TicketNumber combination should be unique. (Currently this is enforced by the application code and I am wondering whether any of the RDBMS support defining this as a DB constraint).

Only a few DBMS engines allow multi-column constraints. All of the DBMS engines that I can think of allow UNIQUE INDEX creation, which would do what you want.

-PatP

only if CustomerID is redundantly carried in the Ticket table

Although there is an applications table, they are probably referring to a join table (customers can have multiple applications) in which the customer id should be carried over to the ticket table unless he has a design problem where applications is not normalized.