Question

I’m using a table to save some text data with this structure:

CREATE TABLE IF NOT EXISTS `strings_en` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `title` varchar(80) COLLATE utf8_unicode_ci NOT NULL,
  `content` text COLLATE utf8_unicode_ci NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB  DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci AUTO_INCREMENT=1;

This table keeps all english “strings” (really this is a text that sometimes is longer that 4kb). When I need a string I call mySQL like this:

SELECT `content` FROM `strings_en` WHERE `id`='42'

Now, I have “internationalized” my application so I have added other tables called “strings_es” or “strings_de” where spanish and dutch languages are used. Each id matches allways the same string in different languages.

In PHP I have a $language variable that contains the current selected language: ‘es’ for spanish, ‘en’ for english, etc. I get an string with this query:

SELECT `content` FROM `strings_$language` WHERE `id`='$string'

The problem is sometimes there is no translation for a given id (this id not exists in the table). For instance, in spanish doesn’t exists id=10 (strings_es), but it exists in english (strings_en). I use *strings_en* as default, so when a string don’t exist in the current language I search in english. Now I’m doing this through a PHP conditional:

$result = query("SELECT `content` FROM `strings_$language` WHERE `id`='$string'");
if (empty($result)) {
    $result = query("SELECT `content` FROM `strings_en` WHERE `id`='$string'");
}

I would like to know how to do this in only one MySQL query.

Answer 1

You should use the language inside your strings table and make it part of the unique key (id+language). This way you won’t have to add new tables when you use more and more languages.

For example:

CREATE TABLE IF NOT EXISTS `strings` (
  `string_id` int(11) NOT NULL AUTO_INCREMENT,
  `id` int(11) NOT NULL ,
  `language` char(2) NOT NULL DEFAULT 'en',
  `title` varchar(80) COLLATE utf8_unicode_ci NOT NULL,
  `content` text COLLATE utf8_unicode_ci NOT NULL,
  PRIMARY KEY (`string_id`) ,
  `i_id_lang`  UNIQUE INDEX (id,language)
) ENGINE=InnoDB  DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci AUTO_INCREMENT=1;

And with

( SELECT * FROM strings WHERE id=? AND language=? UNION 
  SELECT * FROM strings WHERE id=? AND language='en' ) LIMIT 1

you would get a string for the id in a given language or the default language.

Edit: I added a string_id column to act as the unique id.