Re: [Wikitech-l] Use binary to represent varchar rather than UTF8

howard chen wrote:

in all table creation statments, such as

CREATE TABLE `categorylinks` ( `cl_from` int(8) unsigned NOT NULL default '0', `cl_to` varchar(255) binary NOT NULL default '', `cl_sortkey` varchar(255) binary NOT NULL default '', `cl_timestamp` timestamp(14) NOT NULL, UNIQUE KEY `cl_from` (`cl_from`,`cl_to`), KEY `cl_sortkey` (`cl_to`,`cl_sortkey`(128)), KEY `cl_timestamp` (`cl_to`,`cl_timestamp`) ) TYPE=InnoDB;

why use binary to represent varchar, not UTF8?

Brion answered that earlier:

http://lists.wikimedia.org/pipermail/mediawiki-l/2006-February/010267.html

To quote him, utf8 in database will mostly:

* Make indexes larger (3 bytes per character) * Cause failures if you use characters outside the BOM in page titles, usernames, etc.

cheers,