I have an application that deals with businesses and their physical locations. So most tables have a location_id field that logically partitions the data.
One table deals with marketing referrals and each location has a set of sources that they can define in the application. However, ALL locations have an "Unknown" source that is basically a catch-all when the source isn't known. The marketing referral id is the primary auto-incrementing key on the table, and the location_id is a foreign key to the locations table.
As its set up right now, there is an "Unknown" record for each location in the table, each with its corresponding location_id. This means there are a bunch of duplicative records (aside from the location_id). Is this the best possible design for this scenario? I don't really want to remove the foreign key constraint, but in a perfect world I'd see there being a record with an ID of 0 that would serve as the "unknown" record for all the various locations.
Am I overthinking it?