Hacker News new | ask | show | jobs
by hagope 1431 days ago
in that scenario why not use foreign key? any advantage to using constraint?
1 comments

You do both.

The foreign key only gaurantees that the other entity exists.

The unique constraint ensures that only one pair of entities has this relationship, preventing a one-to-many binding.

The distinctness of NULL allows you to have multiple entities with the same NULL value without violating the above UNIQUE constraint.

The "NULL is empty" vs "NULL is unknown" is a series of trade-offs of labor-saving. Imho, the wrong trade-offs were made, but once the choice is made it makes sense to continue and be consistent with it. I'd rather be consistently wrong than inconsistently right.

If the entity doesn't exist, wouldn't it violate the FK and therefore no need for the nullable constraint?