在mysql 5.1中优化查询和表

Question

I have a question regarding optimizing the following query/table

SELECT playitemid,MAX(playdatetime) 
        FROM   buma
        WHERE  licenseid = 1 AND playdatetime > Date_sub(Curdate(), INTERVAL 1 month) 
        GROUP  BY playitemid 

For a table with 11 million records this can sometimes take over 30 seconds.

Here is the create statement for the table.

CREATE TABLE `buma` (
 `bumaid` int(11) NOT NULL AUTO_INCREMENT,
`playitemid` int(11) NOT NULL,
`playdatetime` datetime DEFAULT NULL,
`stopdatetime` datetime DEFAULT NULL, `licenseid` int(11) NOT NULL, editionid` int(11) DEFAULT NULL, 
PRIMARY KEY (`bumaid`),
KEY `ind_buma`(`playdatetime`,`licenseid`,`playitemid`) USING BTREE) ENGINE=InnoDB AUTO_INCREMENT=68644363 DEFAULT CHARSET=latin1;

Is there any way to define a better key or index to speed up the query

Kind regards,

Bjørn

Answer 1

Try to index only playdatetime, licenseid, playitemid alone or groupe only licenseid , playitemid .

MySQL 5.1 is very old (10 years old). Upgrade to 5.7 for a good performance increase (or use a mysql fork like mariadb 10.1).

You can also try to calculate this : Date_sub(Curdate(), INTERVAL 1 month) without mysql (php strtotime or other), and use mysql cache for instant request if you call it more than one time per day, be carefull dont use more than 128MB, it can decrease performances too).