Yes, I agree. There needs to be a check to ensure no other indexes reference the same key.
Sorry to be pedantic David, but it is vital to ensure we are saying the correct thing...
Your response should have read:
I agree. There needs to be a check that no other relationships reference the same named column set.I will almost certainly be teaching grandmother to suck eggs here - but someone at Sparx isn't a grandmother.
We call a column set that is the origin end of a database relationship a Foreign Key because it is native to the current table and because only extremely rarely is it a Key (unique index).
A Foreign Key is a column set that
references a
Key in the
Foreign table.
Actually, Foreign Key is a misnomer - it is actually the Foreign Key
Constraint. Each unique relationship requires a
unique FK Constraint. Because of the Role selection bug described in:
Bug: DB Relationship Roles you can (incorrectly) reselect the same «FK» stereotyped operation.
Consequently, when you delete the relationship, you should be able to (optionally) delete the (unique) «FK» stereotyped operation - but unfortunately, because it's been (incorrectly) reused, you also delete the other relationship's column set!
Paolo
