使用主键的一部分作为外键

Question

I'm building a small MYSQL database, and I currently am having an issue with my entitiy HOME. Its attributes are ROOM_NUM and BUILD_CODE. The two make up a composite primary key, with foreign key BUILD_CODE referencing which building it's located in.

I was wondering if it would be possible to have BUILD_CODE be a single char ('a', 'b', 'c', etc...) and then have a primary key for HOME being something like "A302" or "B205". The first character of that primary key would reference BUILD_CODE, with the remaining numbers being the room number. Then I'd have a unique single value primary key. The reason I ask is so that I don't have to have multiple foreign keys in my PERSON table referencing to HOME.

Answer 1

You may use a surrogate key (a sequence) as the primary key of the entity HOME and Use that as the foreign key of the PERSON. And then for BUILD_CODE and ROOM_NUM declare a unique key to maintain the uniqueness of the combination.

Answer 2

It sounds like HOME needs its own separate identifier. And it doesn't matter what the ID is...as long as it's unique.

Consider the following:

mysql> CREATE TABLE HOME (ID INTEGER NOT NULL auto_increment, ROOM_NUM INTEGER, BUILD_CODE VARCHAR(1), PRIMARY KEY(ID));
Query OK, 0 rows affected (0.14 sec)

mysql> CREATE TABLE PERSON(ID INTEGER NOT NULL auto_increment, NAME VARCHAR(255), HOME_ID INTEGER, PRIMARY KEY(ID), FOREIGN KEY(HOME_ID) REFERENCES HOME(ID));
Query OK, 0 rows affected (0.14 sec)

mysql> INSERT INTO HOME (ROOM_NUM, BUILD_CODE) VALUES (302, "A"), (205, "B");
Query OK, 2 rows affected (0.06 sec)
Records: 2  Duplicates: 0  Warnings: 0

mysql> SELECT * FROM HOME;
+----+----------+------------+
| ID | ROOM_NUM | BUILD_CODE |
+----+----------+------------+
|  1 |      302 | A          |
|  2 |      205 | B          |
+----+----------+------------+
2 rows in set (0.00 sec)

mysql> INSERT INTO PERSON (NAME, HOME_ID) VALUES ("someone", 1);
Query OK, 1 row affected (0.05 sec)

mysql> SELECT * FROM PERSON;
+----+---------+---------+
| ID | NAME    | HOME_ID |
+----+---------+---------+
|  1 | someone |       1 |
+----+---------+---------+
1 row in set (0.00 sec)

mysql> SELECT PERSON.NAME, HOME.BUILD_CODE, HOME.ROOM_NUM FROM HOME JOIN PERSON ON PERSON.HOME_ID = HOME.ID;
+---------+------------+----------+
| NAME    | BUILD_CODE | ROOM_NUM |
+---------+------------+----------+
| someone | A          |      302 |
+---------+------------+----------+
1 row in set (0.00 sec)

You see, it really doesn't matter that HOME has ID values that don't say much about the rest of its data... You're only using that field as an instrument to join the two together .

Of course, you could just as easily have INSERT ed a concatenation of the BUILD_CODE and ROOM_NUM fields into the ID field... That would be fine too. But keep in mind that (a) string lookups use more resources than integer lookups, and (b) it's just one more thing that your application needs to maintain (for example: what happens if you start out by calculating the room numbers as A302, but then after seeing larger room numbers, you decide on A00302? You'd have to go through the entire table--and all its associations--and update the ID . It's not terribly ideal).