I'm working on existing gallery system for which I want to make a few fields multilingual (title, source, description). A distinct table "translations" is used to store these translated texts. The "lang" column of this table stores the language code ("fr", "en", "de"). The "file_id" column stores to which file the translation relates.
I'm looking for the best way in MySQL to output the record for one file, with all its columns, including all translations as columns "title_fr, title_en, ..., description_de.
Tables structure:
CREATE TABLE `files` (
`id` int(10) unsigned NOT NULL auto_increment,
`src` varchar(128) NOT NULL default '',
`name` varchar(128) NOT NULL default '',
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=utf8 AUTO_INCREMENT=1 ;
CREATE TABLE `translations` (
`id` int(10) unsigned NOT NULL,
`file_id` int(10) unsigned NOT NULL,
`lang` char(2) NOT NULL default '',
`title` varchar(255) NOT NULL default '',
`source` varchar(32) default '',
`description` text,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=utf8 AUTO_INCREMENT=1 ;
I tried two ideas, but none worked.
1) The first idea was using VIEWs. Failed because of missing access rights on the shared hosting:
GRANT CREATE VIEW ON *.* TO 'the-user@the-host' IDENTIFIED BY PASSWORD 'here-the-password';
CREATE VIEW French AS
SELECT f.id, title AS title_fr, source AS source_fr, description AS description_fr
FROM files AS f LEFT OUTER JOIN translations AS t ON t.file_id = f.id WHERE (f.id=11 AND t.lang='fr');
CREATE VIEW English AS
SELECT f.id, title AS title_en, source AS source_en, description AS description_en
FROM files AS f LEFT OUTER JOIN translations AS t ON t.file_id = f.id WHERE (f.id=11 AND t.lang='en');
CREATE VIEW German AS
SELECT f.id, title AS title_de, source AS source_de, description AS description_de
FROM files AS f LEFT OUTER JOIN translations AS t ON t.file_id = f.id WHERE (f.id=11 AND t.lang='de');
SELECT * FROM files AS f
LEFT OUTER JOIN French ON f.id=French.id
LEFT OUTER JOIN English ON f.id=English.id
LEFT OUTER JOIN German ON f.id=German.id
GROUP BY f.id;
2) The second idea is inspired from the example in this thread using a pivot table. In my case, I cannot use COUNT() and the following SQL query is not valid, but gives the idea. Sorry as there is some redundancy.
SELECT f.*,
(CASE
WHEN t.`lang`='fr'
THEN t.title
ELSE NULL
) AS title_fr,
(CASE
WHEN t.`lang`='en'
THEN t.title
ELSE NULL
) AS title_en,
(CASE
WHEN t.`lang`='de'
THEN t.title
ELSE NULL
) AS title_de,
(CASE
WHEN t.`lang`='fr'
THEN t.source
ELSE NULL
) AS source_fr,
(CASE
WHEN t.`lang`='en'
THEN t.source
ELSE NULL
) AS source_en,
(CASE
WHEN t.`lang`='de'
THEN t.source
ELSE NULL
) AS source_de
(CASE
WHEN t.`lang`='fr'
THEN t.description
ELSE NULL
) AS description_fr,
(CASE
WHEN t.`lang`='en'
THEN t.description
ELSE NULL
) AS description_en,
(CASE
WHEN t.`lang`='de'
THEN t.description
ELSE NULL
) AS description_de
FROM files AS f WHERE id=11
LEFT OUTER JOIN
translations AS t
ON f.id=t.file_id
GROUP BY id
I also read about CONCAT and CONCAT_WS but they are not what I'm looking for, as I want to make further PHP processing as simple as possible.