This is a discussion issue. Tasks will be defined after this discussion issue is closed.
Currently, our Entity library is pretty simple and primitive. It only supports a single database per program, and there's no distinction between a database and a row. Ideally, we would have a clear distinction between representations of a database, a row, and a collection of rows (eg.: a table), so we could use object-oriented programming to properly represent every single aspect of a database (for now, we'll be focused in databases, collections/tables, and rows/documents).
Below are examples of how I imagine it could be: