Question

I’ve created table Address with this SQL query:

CREATE TABLE `address` (
  `id` bigint(20) NOT NULL AUTO_INCREMENT,
  `Street` varchar(255) COLLATE utf8_unicode_ci DEFAULT NULL,
  `Number` smallint(6) DEFAULT NULL,
  `other_id` bigint(20) NOT NULL,
  PRIMARY KEY (`id`),
  FOREIGN KEY (`other_id`) REFERENCES `other` (`id`) ON DELETE CASCADE ON UPDATE CASCADE
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci;

But there’s also this query:

CREATE TABLE `address` (
  `id` bigint(20) NOT NULL AUTO_INCREMENT,
  `Street` varchar(255) COLLATE utf8_unicode_ci DEFAULT NULL,
  `Number` smallint(6) DEFAULT NULL,
  `other_id` bigint(20) NOT NULL,
  PRIMARY KEY (`id`),
  KEY `other_id` (`other_id`),
  CONSTRAINT `adress_ibfk_1` FOREIGN KEY (`other_id`) REFERENCES `other` (`id`) ON DELETE CASCADE ON UPDATE CASCADE
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci$$

and it seems that booth query create identical tables.

So can anyone explain to me what does this line do:

KEY `other_id` (`other_id`),

and what is the difference between these two lines:

  FOREIGN KEY (`other_id`) REFERENCES `other` (`id`) ON DELETE CASCADE ON UPDATE CASCADE
and
  CONSTRAINT `adress_ibfk_1` FOREIGN KEY (`other_id`) REFERENCES `other` (`id`) ON DELETE CASCADE ON UPDATE CASCADE

If the difference between last two lines is that latter gives name ‘adress_ibfk_1’ to foreign key ? If that’s true – should I do it ? I mean, why should I name foreign keys ? Will I ever need their names ?

Thanks ! 🙂

Answer 1

MySQL interprets KEY as an index, so the second query creates an index on the column other_id.

The difference between the two FK declaration is that you manually set the name in the second line. In the first line, MySQL automatically sets a name.
They do need names, but you don’t necessarily have to be aware of them. Some more advanced RDBMS use them to be more explicit when a query raises an error.