From what I know, if you are documenting a data structure (e.g. JSON) with embedded objects, you might have something like this for the data schema tables:
DISCLAIMER: In the real world, I doubt you'd have a full (e.g.) doctor object embedded in the prescription object. More likely, you'd have a record number for patient, doctor, drug, but the below is useful for this example. :)
You'd then have tables for the aforementioned objects, which themselves could contain objects:
And in turn...
ID NUMBERS OBJECT
Now, to the question. I'm documenting an API where some of the embedded objects contain only one field/object. To use the above example, something like:
PRESCRIPTION OBJECT (as before: contains many fields/objects)
DOCTOR OBJECT (Contains one object)
ID NUMBERS OBJECT (Contains one field)
Per the above, would I be overdoing it to have separate data tables for these objects containing only one field/object? If so, I'd need to find a clean way to embed everything in the parent table.