Furthermore, the technique makes the assumption that the designer understands the nature of foreign keys in conjunction with one-to-many relationships, of many-to-many relationships, and of intersection data.
So in our example we start out with three base tables, each consisting initially of only an entity identifier, as shown in Figure 4a, followed by four other operations to be performed in constructing the relational database: adding simple attribute fields to those tables, adding foreign keys to those tables to satisfy the one-to-many relationships, constructing additional tables to handle the many-to-many relationships, and adding intersection data to those latter tables.
Notice how the simple attributes of the entities are represented by circles attached to the rectangles, and the intersection data are represented by circles attached to the diamond-shaped box of the many-to-many relationship.
Finally, the many-to-many relationship between mechanics and courses is represented by converting its diamond-shaped box into a record type with the intersection data fields brought inside.
Note that the pentagonally represented logical child segments, introduced a part of this process, contain the intersection data fields.
That includes the appropriate insertion of date and grade as intersection data in the logical child segments required as part of the representation of a many-to-many relationship in this kind of structure.
Its nonkey fields are the intersection data fields attached to circles off the diamond-shaped connecting box (DATE and GRADE).