Enforcing "one-to-one-or-more" vs. "one-to-zero-or-more"

Question

Suppose that I have a mini-world consisting of three types of entities, A, B, and C, subject to the following constraints:

1. each A is associated with 0 or more B's, and each B must be associated with exactly one A;
2. each A is associated with 1 or more C, and each C must be associated with exactly one A.

It is easy to come up with a schema that will automatically enforce constraint 1, namely, include in the definition for the B table a foreign key referring to the A table.

The same idea does not quite work for constraint 2, because it does not ensure that at least 1 C is associated with each A.

Is there a standard/canonical way to enforce constraint 2? Or does it depend on the particular DB vendor? If the latter is the case, what would it be for Oracle?

EDIT: Sorry, my original wording of the constraints was incomplete. I fixed this by specifying how many A's must be associated with each B and each C, respectively.

Answer 1

It seems to me that the second half of rule #2 is readily enforced with a foreign-key constraint, as you describe for rule #1.

But the first half of rule #2 seems hard to enforce, in part due to a chicken-vs-egg conflict: Are you going to make 'em add a C before there's an A, so that when you add the A, it will have its (one) C?

But you can't: that new C needs a valid pointer to an A, which doesn't exist yet, and so can't be added.