Data modelling consistency in Relationships

[aanastasiou]

I wonder if it would be possible to take into account the information that is captured in Relationship, RelationshipTo, RelationshipFrom entities at runtime to make things a little bit more efficient and consistent.

Now, it probably only makes sense to maintain one type of directional relationship and that is probably RelationshipTo.

RelationshipTo makes sense from a data modelling point of view since it implies attaching an attribute to the specific entity it belongs to. RelationshipFrom seems a little bit strange because, why would I establish an attribute on the entity I am modelling that really, belongs to the entity it is coming from (?). It is also not a good practice to establish reciprocal relationships (a pair of From,To) where you would only need one relationship without any indication of direction.

Which brings us to the second point, about using Relationship which is probably a little bit more demanding.

When I am establishing a Relationship, since it is not directional, I would expect an attribute to 'pop-up' in both entities at runtime. For example, if I am modelling something like COUNTRY--BELONGS--CITY, I would establish a relationship belongs = Relationship("Country", "BELONGS") on the City, but, I would also expect the Country to now have an attribute at runtime that allows me to traverse the data structure in exactly the same way as it is implied by the schema.

Altenatively, I wonder if it would be possible to retain both RelationshipTo, RelationshipFrom but only for the purpose of allowing the user to explicitly specify that they are after a reciprocal connection and not a directional one. Therefore, directional ones would still have to be specified by a RelationshipTo declaration.