Question

I have the following 3 tables

CREATE TABLE BUBBLES_CSC (
CSC_ID integer UNSIGNED NOT NULL AUTO_INCREMENT,
COUNTRY_NAME varchar(255) NOT NULL,
STATE_NAME varchar(255) NOT NULL,
CITY_NAME varchar(255) NOT NULL,
PRIMARY KEY(CSC_ID,STATE_NAME,CITY_NAME),
INDEX(CSC_ID ,CITY_NAME,COUNTRY_NAME )
);

CREATE TABLE BUBBLES_REGION (
REGION_ID integer UNSIGNED NOT NULL AUTO_INCREMENT,
REGION_NAME varchar(255) NOT NULL,
CSC_ID integer UNSIGNED NOT NULL,
PRIMARY KEY(REGION_ID),
FOREIGN KEY (CSC_ID) REFERENCES BUBBLES_CSC(CSC_ID) ON UPDATE CASCADE ON DELETE CASCADE,
INDEX(REGION_ID ,REGION_NAME ,CSC_ID )
);

CREATE TABLE BUBBLES_HOTEL (
HOTEL_ID integer UNSIGNED NOT NULL AUTO_INCREMENT,
NAME varchar(255) NOT NULL,
DESCRIPTION varchar(255) DEFAULT "No description",
CSC_ID integer UNSIGNED NOT NULL,
REGION_ID integer UNSIGNED NOT NULL,
STREET_ADDRESS varchar(255) DEFAULT NULL,
PRIMARY KEY(HOTEL_ID),
FOREIGN KEY (CSC_ID) REFERENCES BUBBLES_CSC(CSC_ID) ON UPDATE CASCADE,
FOREIGN KEY (REGION_ID) REFERENCES BUBBLES_REGION(REGION_ID) ON UPDATE CASCADE,
INDEX(HOTEL_ID,NAME,CSC_ID,REGION_ID)
)CHARACTER SET=utf8;

Now i know how to do a foreign key, But how to specify that the CSC_ID in BUBBLES_HOTEL should be same as the CSC_ID specified in BUBBLES_REGION for the row pointed to by REGION_ID in BUBBLES_HOTEL

Answer 1

The proper way to do this is not to have the column CSC_ID in BUBBLES_HOTEL at all. Each row in BUBBLES_HOTEL contains a link to a row in BUBBLES_REGION, which in turn contains a link to a row in BUBBLES_CSC. That’s how you identify to which row in BUBBLES_CSC a given row in BUBBLES_HOTEL corresponds. Your current design is not normalised.

Having said that, you could achieve what you ask by replacing the two foreign keys on BUBBLES_HOTEL with just one:

FOREIGN KEY (CSC_ID, REGION_ID) REFERENCES BUBBLES_REGION (CSC_ID, REGION_ID)