Question

This query is working perfectly on my dev machine, a PC with mysql 5.1.53

 SELECT DISTINCT * 
 FROM   posts P1
 WHERE  user_id IN (2,1000001) AND NOT track_id = 34 AND
        (SELECT COUNT(*) FROM posts P2 
            WHERE P2.user_id = P1.user_id AND P2.id > P1.id AND P2.track_id <> 34)
         <= 1
 GROUP BY 
        track_id 
 ORDER BY 
        id desc LIMIT 5

when i run the very same piece of code in the same database on my live server (debian, mysql 5.1.39), i get this error:

Unknown column 'P1.user_id' in 'where clause': 

How could this be? Any ideas?

Result of show create table posts on the server

CREATE TABLE `posts` (
`id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`track_id` int(11) unsigned DEFAULT '0',
`tag_id` int(11) unsigned DEFAULT NULL,
`user_id` int(11) DEFAULT '0',
`comment` tinytext,
`created_at` datetime DEFAULT NULL,
`commentcount` tinyint(11) unsigned DEFAULT '0',
PRIMARY KEY (`id`),
KEY `track_id` (`track_id`),
KEY `tag_id` (`tag_id`),
KEY `user_id` (`user_id`)
) ENGINE=MyISAM AUTO_INCREMENT=484 DEFAULT CHARSET=utf8 |

Answer 1

As a lot of comments suggested, it was an issue with case sensitivity. In the SQL query on the server, the alias of posts was set to “p1”, but the subselect was looking for “P1”. Mistake.