Question

表结构如下：

CREATE TABLE `crm_member` (
 `member_id` int(11) NOT NULL AUTO_INCREMENT,
 `shop_id` int(11) NOT NULL,
 `nick` varchar(255) NOT NULL DEFAULT '',
 `name` varchar(255) NOT NULL DEFAULT '',
 `mobile` varchar(255) NOT NULL DEFAULT '',
 `grade` int(11) NOT NULL DEFAULT '-1',
 `trade_count` int(11) NOT NULL,
 `trade_amount` float NOT NULL,
 `last_trade_time` int(11) NOT NULL,
 `trade_from` tinyint(4) NOT NULL,
 `avg_price` float NOT NULL,
 `seller_flag` tinyint(1) NOT NULL,
 `is_black` tinyint(1) NOT NULL DEFAULT '0',
 `created` int(11) NOT NULL,
 PRIMARY KEY (`member_id`),
 UNIQUE KEY `shop_id` (`shop_id`,`nick`),
 KEY `last_trade_time` (`last_trade_time`),
 KEY `idx_shop_id_grade` (`shop_id`,`grade`),
 KEY `idx_shopid_created` (`shop_id`,`created`),
 KEY `idx_trade_amount` (`shop_id`,`trade_amount`),
 KEY `idx_trade_count` (`shop_id`,`trade_count`)
) ENGINE=InnoDB  DEFAULT CHARSET=utf8

该表在shop_id 3498706下具有2148037行；

我的查询就像

SELECT AVG(trade_count) as trade_count, AVG(trade_amount) as trade_amount, AVG(grade) as grade from `crm_member_0141` WHERE shop_id = '3498706' and grade >= 0 and trade_count > 0 and is_black = 0 LIMIT 1

查询执行大约30秒。

解释结果

mysql> explain SELECT member_id, AVG(trade_count) as trade_count, AVG(trade_amount) as trade_amount, AVG(grade) as grade from `crm_member_0141` WHERE shop_id = 3498706 and grade >= 0 and trade_count > 0 and is_black = 0 order by member_id LIMIT 1;
+----+-------------+-----------------+------------+------+-------------------------------------------------------------------------------+---------+---------+-------+---------+----------+-------------+
| id | select_type | table           | partitions | type | possible_keys                                                                 | key     | key_len | ref   | rows    | filtered | Extra       |
+----+-------------+-----------------+------------+------+-------------------------------------------------------------------------------+---------+---------+-------+---------+----------+-------------+
|  1 | SIMPLE      | crm_member_0141 | NULL       | ref  | shop_id,idx_shop_id_grade,idx_shopid_created,idx_trade_amount,idx_trade_count | shop_id | 4       | const | 1074018 |     1.11 | Using where |
+----+-------------+-----------------+------------+------+-------------------------------------------------------------------------------+---------+---------+-------+---------+----------+-------------+

如何加快查询速度？

Answer 1

尽管shop_id, grade可能会加快行的选择速度，但看起来mysql选择使用主键而不是shop_id, grade索引。

您可以使用USE INDEX指令告诉mysql使用特定的索引进行查询：

尝试在表名后的查询中添加USE INDEX (idx_shop_id_grade)，以查看计算是否更快。

否则，如果此查询特别有用并且经常在您的应用中调用，则可以为此查询构建更专业的索引：

您的查询选择shop_id和is_black的特定值，然后对grade和trade_count进行范围选择。

我建议尝试使用shop_id, is_black, grade, trade_count进行索引。

注意：显然，首先要在测试数据库上进行测试

如何加快mysql查询计数平均值？

1 个答案: