Very slow subquery in MySQL with GROUP BY

Question

I've got this SQL:

SELECT COUNT( d.id ) AS sum, d.user_agent
FROM debug AS d
WHERE d.id IN (
    SELECT MAX(d2.id)
    FROM debug AS d2
    GROUP BY d2.user_id
)
GROUP BY d.user_agent

I want to get the amounts of similar useragents but only the useragents of the last entry of each user.

=> COUNT(d.id)

=> GROUP BY d2.user_id

I tried both SQLs separated and wrote the response of the subquery plain in the other query. The execution time was far under a second. Both together about 25 seconds.

My table structure:

CREATE TABLE `debug` (
  `id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `event` varchar(127) NOT NULL,
  `user_id` int(11) NOT NULL,
  `user_agent` varchar(255) NOT NULL,
  `timestamp` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
  `parameters` varchar(5000) NOT NULL,
  PRIMARY KEY (`id`),
  KEY `user_id` (`user_id`),
  KEY `user_agent` (`user_agent`)
) ENGINE=InnoDB  DEFAULT CHARSET=utf8;

SQL-Fiddle: Here

Any ideas how I can optimize it?

Answer 1

Is this any better?

SELECT x.user_agent
     , COUNT(*) total
  FROM debug x
  JOIN 
     ( SELECT user_id
            , MAX(id) max_id
         FROM debug 
        GROUP
           BY user_id
     ) y
    ON y.user_id = x.user_id
   AND y.max_id = x.id
 GROUP BY user_agent;