0316 【案例】MySQL count操作优化案例一则

 

转自http://blog.itpub.net/22664653/viewspace-1791124/ 

一 背景
 某业务的数据库定期报 thread_runing 飙高,通定位发现一个慢查询sql导致会话堆积。执行sql 耗时如下 

root@db 05:32:05>select count(item_id) from xxxtable where selid = 345705650 and end_time > now();
+----------------+
| count(item_id) |
+----------------+
| 2247052 |
+----------------+
1 row in set (4.65 sec)

二 分析  
慢查询表结构如下 

root@db >show create table xxxtable G
*************************** 1. row ***************************
       Table: uac_shop_item_promotion_0091
Create Table: CREATE TABLE `uac_shop_item_promotion_0091` (
  `id` bigint(20) unsigned NOT NULL AUTO_INCREMENT COMMENT '主键',
  `gmt_modified` datetime NOT NULL COMMENT '修改时间',
  `selid` bigint(20) NOT NULL COMMENT '分表字段',
  `end_time` datetime NOT NULL COMMENT '活动结束时间',
  `item_id` bigint(20) NOT NULL COMMENT '商品id',
  PRIMARY KEY (`id`),
  UNIQUE KEY `idx_uq_item` (`item_id`),
  KEY `idx_deller_id_end_time` (`selid`,`end_time`),
  KEY `idx_deller_id_start_time` (`selid`,`start_time`),
  KEY `idx_seller_item_start` (`selid`,`start_time`,`item_id`)
) ENGINE=InnoDB AUTO_INCREMENT=42132149 DEFAULT CHARSET=gbk COMMENT='索引表'
1 row in set (0.00 sec)

很明显出现问题的sql由于使用了count(item_id) ,而item_id字段并没有和 selid 和end_time 构成有效索引  故该sql 没有合理的使用索引 。查看其直系计划

root@db >explain select count(item_id) from xxxtable
        >where selid = 345705650 and end_time > now() G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: xxxtable
         type: ref
possible_keys: idx_deller_id_end_time,idx_deller_id_start_time,idx_seller_item_start
          key: idx_deller_id_end_time
      key_len: 8
          ref: const
         rows: 1726757
        Extra: Using where
1 row in set (0.00 sec)

从key_len=8 和Extra: Using where 可以看出MySQL没有完全利用到idx_deller_id_end_time组合索引而是利用到了 selid字段作为过滤条件回表查询。
count(item_id)的意思是符合where条件的结果集中item_id非空集合的总和。
三 如何优化
根据该sql的业务需求是需要获取到某商家参加活动且活动截止时间大于当前时间的商品总数,可以使用如下sql满足要求:

select count(*) from xxxtable where selid = 345705650 and end_time > now()

执行时间仅为原来的1/4,新的sql发布之后thread_running报警消失，业务校验时间明显缩短。

root@db >select count(*) from xxxtable where selid = 345705650 and end_time > now();
+----------+
| count(*) |
+----------+
| 2247052 |
+----------+
1 row in set (0.82 sec)
root@db >select count(1) from xxxtable where selid = 345705650 and end_time > now();
+----------+
| count(1) |
+----------+
| 2247052 |
+----------+
1 row in set (0.79 sec)

优化后的sql的explain 方式如下:

root@db >explain select count(*) from xxxtable where selid = 345705650 and end_time > now() G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: xxxtable
         type: range
possible_keys: idx_deller_id_end_time,idx_deller_id_start_time,idx_seller_item_start
          key: idx_deller_id_end_time
      key_len: 16
          ref: NULL
         rows: 1726768
        Extra: Using where; Using index
1 row in set (0.00 sec)

四 小结
 a 这个问题是在没有修改索引的基础中做出的优化,老的sql没有有效的利用当前的索引导致耗时操作
 b 对于不同count类型的sql 总结如下
   count(*)/count(1) 返回结果集的总和包括null和重复的值。
   count(column) 返回结果集中非空 column 的总和,执行查询的过程中会校验字段是否非空。
 c 在业务设计的时候 满足业务逻辑的前提下推荐使用count(*).
 d 从官方文档中摘录 Using where 和 Using index 的区别 

Using index
 The column information is retrieved from the table using only information in the index tree without having to do an additional seek to read the actual row. This strategy can be used when the query uses only columns that are part of a single index.
 If the Extra column also says Using where, it means the index is being used to perform lookups of key values. Without Using where, the optimizer may be reading the index to avoid reading data rows but not using it for lookups. For example, if the index is a covering index for the query, the optimizer may scan it without using it for lookups. For InnoDB tables that have a user-defined clustered index, that index can be used even when Using index is absent from the Extra column. This is the case if type is index and key is PRIMARY.
 Using where
 A WHERE clause is used to restrict which rows to match against the next table or send to the client. Unless you specifically intend to fetch orexamine all rows from the table, you may have something wrong in your query if the Extra value is not Using where and the table join type is ALLor index. Even if you are using an index for all parts of a WHERE clause, you may see Using where if the column can be NULL.