I'm researching OData as a RESTful interface for a database. The data is structured in a very unusual way and normal tables and rows do not apply, in fact, some stuff just exists in in-memory collections and objects.
Can I build my own arbitrary mapping system between the entities that make up 'feeds' and the sources behind, this might mean aggregating from sources and building the entities on the fly?
I'm just looking for yes/no (why not) and maybe some pointers to relevant reading material.
Many thanks
Luke