MySQLcount(*)之索引选择
覆盖索引对于一些统计问题,如下:
创新互联主要从事网站建设、做网站、网页设计、企业做网站、公司建网站等业务。立足成都服务江城,十余年网站建设经验,价格优惠、服务专业,欢迎来电咨询建站服务:18982081108
MySQL > show create table test1 \G *************************** 1. row *************************** Table: test1 Create Table: CREATE TABLE `test1` ( `id` bigint(16) NOT NULL AUTO_INCREMENT, `order_seq` bigint(16) NOT NULL, PRIMARY KEY (`id`), KEY `idx_id` (`id`), KEY `idx_id_ordseq` (`id`,`order_seq`) ) ENGINE=InnoDB AUTO_INCREMENT=15002212 DEFAULT CHARSET=utf8mb4 1 row in set (0.00 sec) MySQL > explain select count(*) from test1 where id>10000 and id<20000; +----+-------------+-------+------------+-------+------------------------------+--------+---------+------+------+----------+--------------------------+ | id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra | +----+-------------+-------+------------+-------+------------------------------+--------+---------+------+------+----------+--------------------------+ | 1 | SIMPLE | test1 | NULL | range | PRIMARY,idx_id,idx_id_ordseq | idx_id | 8 | NULL | 9999 | 100.00 | Using where; Using index | +----+-------------+-------+------------+-------+------------------------------+--------+---------+------+------+----------+--------------------------+ 1 row in set, 1 warning (0.00 sec)
innodb存储引擎选择了id字段的辅助索引,而不是聚集索引来统计,更不是联合索引。原因是辅助索引远小于聚集索引,选择辅助索引可以减少IO资源消耗。
而另外一个统计场景:
select count(*) from test1 where order_seq > 1502131212577 and order_seq< 202007080947244761;
test1表建有id和 order_seq 字段的联合索引。
MySQL > show create table test1 \G *************************** 1. row *************************** Table: test1 Create Table: CREATE TABLE `test1` ( `id` bigint(16) NOT NULL AUTO_INCREMENT, `order_seq` bigint(16) NOT NULL, PRIMARY KEY (`id`), KEY `idx_id` (`id`), KEY `idx_id_ordseq` (`id`,`order_seq`) ) ENGINE=InnoDB AUTO_INCREMENT=15002212 DEFAULT CHARSET=utf8mb4 1 row in set (0.00 sec) MySQL > explain select count(*) from test1 where order_seq > 1502131212577 and order_seq< 202007080947244761; +----+-------------+-------+------------+-------+---------------+---------------+---------+------+----------+----------+--------------------------+ | id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra | +----+-------------+-------+------------+-------+---------------+---------------+---------+------+----------+----------+--------------------------+ | 1 | SIMPLE | test1 | NULL | index | NULL | idx_id_ordseq | 16 | NULL | 15068082 | 11.11 | Using where; Using index | +----+-------------+-------+------------+-------+---------------+---------------+---------+------+----------+----------+--------------------------+ 1 row in set, 1 warning (0.00 sec)
这里使用条件 order_seq 查询,一般情况下使用不了联合索引的,但是这个案例中的查询,利用到覆盖索引的信息。possible_keys依然为null,但是key是idx_id_ordseq,extra里出现Using index,表示为覆盖索引。
本文标题:MySQLcount(*)之索引选择
文章来源:http://azwzsj.com/article/jggscd.html