当一个表引用多个表时的关系数据库设计

Question

For exemple let's imagine an " appointments " table that keeps track of several Types of appointments (eye, teeth ..) This table could have an ID, date, doctor_id, Appointment_type and other columns that are commun to all appointment types. Then each type of appointment will has it's own table to hold data specific to this appointements type.

I chose this design as we have a significant amount of data commun to all appointments so i don't want to repeat it and each appointments holds very different type of data, so having only one appointent table that has lots of NULL field seemed unhandy.

Knowing that I will mostly need to query for all appointment type at once what would be the best way to organize these tables ?

Answer 1

There're two approaches:

One table per hierarchy (aka Single Table Inheritance) You put all the attributes of all appointment types in one table and use a "type" filed" to distinguish the type of appointments. This is fast because there's no join. But this table may have many columns.

One table per subclass (aka Multiple Table Inheritance) This way you have a base class table including the common fields, and use polymorphic association to link to the records in the subclass tables. This is slower, requires more tables. Joins can be difficult to write if you want to query all appointment types.

I think for your application (query all appointment types), one table per hierarchy is better.