I would like to model something similar to the following scenario in a database built on MySQL:
I am managing reviews of texts. These texts can either be books or parts of books (e. g. different articles). There may be multiple reviews per text. It can also be guaranteed there is an identifying relationship between from an article to the containing book.
My approach was the following:
The table Review
has an ID
as a primary key, a name
and a foreign key to the primary keys of Book
and Book Part
. Book
and Book Part
have primary keys and some custom attributes for each (see below). Additionally, there is an identifying 1-to-n relationship between Book
and Book Part
.
Is this a viable approach?
Edit: I created a picture for the model.
Description of attributes
Review
------
author - author of the review
rating - an integer describing how well the author liked the text
idBook - book the review is about
idBook_Part - book part the review is about
Book
----
author - author of the book
year - year the book was published
Book Part
---------
author - author of the book part
start page - page where the book part starts in the book
end page - page where the book part ends in the book
idBook - book that contains the part
Note: The attributes of Book and Book Part are intentionally not supersets of each other.
To make sure Review
contains either an idBook
or an idBook_Part
I have written a MySQL trigger for INSERT
and UPDATE
statements that checks that exactly one of them is NULL
.
References to literature are appreciated as well.