Re: select distinct slower in 5.0 than 4.1

[Saar Barhoom]

CREATE TABLE `bucket_lastname` (
`PID` int(10) unsigned NOT NULL default '0',
`lastname` char(32) default NULL,
UNIQUE KEY `BUCKET_lastname_CONST` (`PID`,`lastname`),
KEY `BUCKET_lastname_ATT` TYPE BTREE (`lastname`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1;


explain select distinct(pid) from bucket_lastname where lastname = 'Smith'

Result with 4.1.12
1,"SIMPLE","bucket_lastname","ref","BUCKET_lastname_ATT","BUCKET_lastname_ATT",33,"const",709,"Using where; Using temporary"

Result with 5.0.16
1,"SIMPLE","bucket_lastname","range","BUCKET_lastname_ATT","BUCKET_lastname_CONST","",8,"Using where; Using index for group-by"