http://blogs.mysql.com/peterg/2008/06/09/truncate-now-and-unto-ages-of-ages/ By MySQL development-department workers Wed, 25 Nov 2009 01:20:10 +0000 http://wordpress.org/?v=2.6.1 http://blogs.mysql.com/peterg/2008/06/09/truncate-now-and-unto-ages-of-ages/#comment-195 Arjen Lentz Mon, 09 Jun 2008 04:40:57 +0000 http://blogs.mysql.com/peterg/2008/06/09/truncate-now-and-unto-ages-of-ages/#comment-195 The exact behaviour of TRUNCATE will depend on the storage engine used, and the version of MySQL. MyISAM has "always" supported truncate by truncating MYD/MYI and rewriting the MYI file header. Internally it's not quite a drop, but the reinitialisation the MYI header does reset the auto-inc counter. By default, engines would map it to DELETE FROM table, making it a DML query with all the bells, and no resetting of auto-inc. Somewhere in the 4.0 line, InnoDB started supporting TRUNCATE the way MyISAM does. So, despite all your proof, I'd say "it depends" (on the engine, mostly). The exact behaviour of TRUNCATE will depend on the storage engine used, and the version of MySQL.
MyISAM has “always” supported truncate by truncating MYD/MYI and rewriting the MYI file header. Internally it’s not quite a drop, but the reinitialisation the MYI header does reset the auto-inc counter.
By default, engines would map it to DELETE FROM table, making it a DML query with all the bells, and no resetting of auto-inc.
Somewhere in the 4.0 line, InnoDB started supporting TRUNCATE the way MyISAM does.
So, despite all your proof, I’d say “it depends” (on the engine, mostly).

]]>