我需要优化MySQL查询,这需要花费大量时间来加载。
这是:
SELECT
p.id,
UNIX_TIMESTAMP(p.last_answer_date) AS last_answer_date_timestamp,
p.sender_id,
p.recipient_id,
p.is_read_sender,
p.last_answer_user_id,
p.is_read_recipient,
(SELECT m.read FROM pm_message m WHERE m.conv_id = p.id AND m.user_id != $user_id ORDER BY m.date DESC LIMIT 1) AS read_status,
(SELECT m.content FROM pm_message m WHERE m.conv_id = p.id ORDER BY m.date DESC LIMIT 1) AS last_message,
(SELECT u.username FROM user u WHERE (u.id = p.sender_id OR u.id = p.recipient_id) AND u.id != $user_id LIMIT 1) AS from_username,
(SELECT u.id FROM user u WHERE (u.id = p.sender_id OR u.id = p.recipient_id) AND u.id != $user_id LIMIT 1) AS from_userid,
(SELECT ui.gender FROM user_info ui WHERE (ui.user_id = p.sender_id OR ui.user_id = p.recipient_id) AND ui.user_id != $user_id LIMIT 1) AS from_gender,
(SELECT ph.thumb_url FROM photo ph, user_info ui WHERE ui.main_photo = ph.id AND (ph.user_id = p.sender_id OR ph.user_id = p.recipient_id) AND ph.user_id != $user_id LIMIT 1) AS from_thumb_url
FROM pm_conv p
WHERE p.sender_id = $user_id OR p.recipient_id = $user_id
ORDER BY p.last_answer_date DESC LIMIT 25;
这个查询得到了我想要的结果,但它真的很慢......而且我认为嵌套选择是这个查询速度太慢的原因。
以下是此查询的表结构:
CREATE TABLE IF NOT EXISTS `photo` (
`id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`url` varchar(255) DEFAULT NULL,
`thumb_url` varchar(255) DEFAULT NULL,
`user_id` int(11) NOT NULL,
`date` datetime NOT NULL,
`status` int(11) NOT NULL,
`votes` int(11) DEFAULT '0',
`comments` int(11) DEFAULT '0',
`views` int(11) DEFAULT '0',
`text` text,
PRIMARY KEY (`id`),
KEY `user_id` (`user_id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=1 ;
CREATE TABLE IF NOT EXISTS `pm_conv` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`date` datetime NOT NULL,
`sender_id` int(11) NOT NULL,
`recipient_id` int(11) NOT NULL,
`last_answer_date` datetime NOT NULL,
`nb_messages` int(11) NOT NULL,
`is_read_sender` int(11) NOT NULL,
`is_read_recipient` int(11) NOT NULL DEFAULT '0',
`last_answer_user_id` int(11) DEFAULT NULL,
PRIMARY KEY (`id`),
KEY `recipient_id` (`recipient_id`),
KEY `sender_id` (`sender_id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=1 ;
CREATE TABLE IF NOT EXISTS `pm_message` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`date` datetime NOT NULL,
`content` text NOT NULL,
`user_id` int(11) NOT NULL,
`conv_id` int(11) NOT NULL,
`read` int(11) DEFAULT '0',
PRIMARY KEY (`id`),
KEY `user_id` (`user_id`),
KEY `conv_id` (`conv_id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=1 ;
CREATE TABLE IF NOT EXISTS `user` (
`id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`encrypt_id` varchar(255) DEFAULT NULL,
`register_date` datetime DEFAULT NULL,
`last_login_date` datetime DEFAULT NULL,
`username` varchar(255) DEFAULT NULL,
`password` varchar(255) DEFAULT NULL,
`email` varchar(255) DEFAULT NULL,
`banned` int(11) DEFAULT NULL,
`banned_reason` text,
`first_step_form` int(11) DEFAULT '0',
`status` int(11) DEFAULT NULL,
`valid_snapchat` int(11) DEFAULT '0',
`introduced_forum` int(11) DEFAULT '0',
`referer` varchar(255) DEFAULT NULL,
`allow_social_featuring` int(11) DEFAULT NULL,
`rank` int(11) DEFAULT NULL,
`fb_id` bigint(20) DEFAULT NULL,
`rate_app_status` int(11) DEFAULT NULL,
`last_activity_date` datetime DEFAULT NULL,
PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=1 ;
CREATE TABLE IF NOT EXISTS `user_info` (
`id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`gender` int(11) DEFAULT NULL,
`birthday` date DEFAULT NULL,
`about` text,
`main_photo` int(11) DEFAULT NULL,
`country` varchar(100) DEFAULT NULL,
`city` varchar(100) DEFAULT NULL,
`relation_type` varchar(30) DEFAULT NULL,
`user_id` int(11) DEFAULT NULL,
`fb_link` varchar(255) DEFAULT NULL,
`twitter_link` varchar(255) DEFAULT NULL,
`youtube_link` varchar(255) DEFAULT NULL,
`instagram_link` varchar(255) DEFAULT NULL,
`app_pref_forum` int(11) DEFAULT NULL,
`app_pref_pm` int(11) DEFAULT NULL,
`app_pref_snapchat_request` int(11) DEFAULT NULL,
`browse_invisibly` int(11) DEFAULT '0',
PRIMARY KEY (`id`),
KEY `user_id` (`user_id`),
KEY `main_photo` (`main_photo`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=1 ;
有人可以帮我优化这个繁重的查询吗?
谢谢!
答案 0 :(得分:0)
您可以在解释计划中看到一些表正由低效索引访问。尝试计算所有表的统计数据,以查看它是否发生了变化(使用分析表)。
答案 1 :(得分:0)
您可以加入user
表,以便一次获取用户名和ID,而不是有两个子查询,可能您可以对pm_message
执行相同操作,但由于子查询具有不同的条件,因此它有点棘手。
我还将user
和user_info
表组合在一起,因为我可以看到它们具有一对一的关系,因此将这些数据存储在不同的表中是没有意义的。这将允许你摆脱第4个子查询并简化第5个子查询
在某些情况下,最好执行多个查询,而不是使用子查询执行查询。