This article addresses, in general, the process of converting your WordPress MySQL database tables from one character set to another. Warning: character set conversion is not a simple process. Please complete a backup of your database before attempting any conversion.
Character set and collation can now be defined
Beginning with Version 2.2, WordPress allows the user to define both the database character set and the collation in their wp-config.php file. Setting the DB_CHARSET and DB_COLLATE values in wp-config.php causes WordPress to create the database with the appropriate settings. But, the setting can only be designated for new installations, not for 'already installed' copies of WordPress. The rest of this article will explain how to convert the character set and collation for existing WordPress installations.
Converting your database
Before beginning any conversion, please backup your database. Backing Up Your Database has easy to following instructions.
For discussion purposes, it is assumed you have a database in the latin1 character set that needs converting to a utf8 character set.
To convert character sets requires using the the MySQL ALTER TABLE command. When converting the character sets, all TEXT (and similar) fields are converted to UTF-8, but that conversion will BREAK existing TEXT because the conversion expects the data to be in latin1, but WordPress may have stored unicode characters in a latin1 database, and as a result, data could end up as garbage after a conversion!
The solution is to ALTER all TEXT and related fields to their binary counterparts, then alter the character set and finally change the binary data type fields back to TEXT.
- Place notice that blog is out of service
- Backup database
- ALTER TABLE wp_users MODIFY display_name BLOB;
- ...ALTER TABLE commands for all other tables/columns...
- ALTER DATABASE wordpress charset=utf8;
- ALTER TABLE wp_users charset=utf8;
- ...ALTER TABLE command for all other tables...
- ALTER TABLE wp_users MODIFY display_name TEXT CHARACTER SET utf8;
- ...ALTER TABLE for all other tables/columns...
- Add DB_CHARSET and DB_COLLATE definitions to wp-config.php
- Place blog back on-line
The string field types need to be converted to their binary field types counterparts. The list is as follows:
- CHAR -> BINARY
- VARCHAR -> VARBINARY
- TINYTEXT -> TINYBLOB
- TEXT -> BLOB
- MEDIUMTEXT -> MEDIUMBLOB
- LONGTEXT -> LONGBLOB
This information was originally posted by member g30rg3x in Forum Thread 117955.
ENUM and SET have more specific conversion rules: Set the character set to binary, or to UTF8 if you are sure that no ENUM or SET field has special characters that might get garbled during conversion. The SQL for this is:
- ALTER TABLE wp_links CHANGE link_visible link_visible ENUM('Y','N') CHARACTER SET utf8;
The field name does need to be repeated, as well as the ENUM specification.
When specifying BINARY and VARBINARY, the field length also needs to be specified, and needs to be the same value as the original CHAR and VARCHAR field length. In other words, VARCHAR(200) becomes VARBINARY(200).
So, in Steps 3 and 4 change CHAR, VARCHAR, TEXT, ENUM, and SET fields to their binary counterparts (BLOB, VARBINARY, etc), in Step 5 switch the database to utf8, in Steps 6 and 7 switch all the tables to utf8, and finally, in Steps 8 and 9 return the binary fields back to the respective CHAR, VARCHAR, TEXT, ENUM, and SET data types with the utf8 character set.
The key to the conversion is that a field with a binary data type, unlike CHAR, VARCHAR, TEXT, ENUM, and SET fields, will not be converted to garbage when the database and tables are switched to utf8.
Conversion Scripts and Plugins
In the WordPress Forums, Member andersapt, in Forum Thread 117955 submitted a conversion script, Convert UTF8 SQL Generator, to automatically convert a WordPress database. (This link is currently dead.)
A plugin, UTF-8 Database Converter, is available from g30rg3_x. Carefully review the readme file included with the plugin.
Discussions on character sets
- Character set at Wikipedia
- Unicode at Wikipedia
- UTF-8 at Wikipedia
- Character sets and collation at MySQL
- Character Sets and Collations That MySQL Supports
- Gentoo tip on converting latin1 to utf8 in MySQL
- Alex King's blog about latin1 to utf8 conversion