MySQL Explain Type列

对于MySQL执行计划的获取，我们可以通过explain方式来查看，explain方式看似简单，实际上包含的内容很多，尤其是输出结果中的type类型列。理解这些不同的类型，对于我们SQL优化举足轻重，本文仅描述explian输出结果中的type列，同时给出其演示。

一、EXPLAIN 语句中type列的值

type：
    连接类型
    system          表只有一行
    const           表最多只有一行匹配，通用用于主键或者唯一索引比较时
    eq_ref          每次与之前的表合并行都只在该表读取一行，这是除了system，const之外最好的一种，
                    特点是使用=，而且索引的所有部分都参与join且索引是主键或非空唯一键的索引
    ref             如果每次只匹配少数行，那就是比较好的一种，使用=或<=>，可以是左覆盖索引或非主键或非唯一键
    fulltext        全文搜索
    ref_or_null     与ref类似，但包括NULL
    index_merge     表示出现了索引合并优化(包括交集，并集以及交集之间的并集)，但不包括跨表和全文索引。
                    这个比较复杂，目前的理解是合并单表的范围索引扫描（如果成本估算比普通的range要更优的话）
    unique_subquery 在in子查询中，就是value in (select...)把形如“select unique_key_column”的子查询替换。
                    PS：所以不一定in子句中使用子查询就是低效的！
    index_subquery  同上，但把形如”select non_unique_key_column“的子查询替换
    range           常数值的范围
    index           a.当查询是索引覆盖的，即所有数据均可从索引树获取的时候（Extra中有Using Index）；
                    b.以索引顺序从索引中查找数据行的全表扫描（无 Using Index）；
                    c.如果Extra中Using Index与Using Where同时出现的话，则是利用索引查找键值的意思；
                    d.如单独出现，则是用读索引来代替读行，但不用于查找
    all             全表扫描

二、连接类型部分示例

1、all
-- 环境描述
(root@localhost) [sakila]> show variables like 'version';
+---------------+--------+
| Variable_name | Value  |
+---------------+--------+
| version       | 5.6.26 |
+---------------+--------+

MySQL采取全表遍历的方式来返回数据行，等同于Oracle的full table scan
(root@localhost) [sakila]> explain select count(description) from film;
+----+-------------+-------+------+---------------+------+---------+------+------+-------+
| id | select_type | table | type | possible_keys | key  | key_len | ref  | rows | Extra |
+----+-------------+-------+------+---------------+------+---------+------+------+-------+
|  1 | SIMPLE      | film  | ALL  | NULL          | NULL | NULL    | NULL | 1000 | NULL  |
+----+-------------+-------+------+---------------+------+---------+------+------+-------+

2、index
MySQL采取索引全扫描的方式来返回数据行，等同于Oracle的full index scan
(root@localhost) [sakila]> explain select title from film \G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: film
         type: index
possible_keys: NULL
          key: idx_title
      key_len: 767
          ref: NULL
         rows: 1000
        Extra: Using index
1 row in set (0.00 sec)

3、  range
索引范围扫描，对索引的扫描开始于某一点，返回匹配值域的行，常见于between、<、>等的查询
等同于Oracle的index range scan
(root@localhost) [sakila]> explain select * from payment where customer_id>300 and customer_id<400\G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: payment
         type: range
possible_keys: idx_fk_customer_id
          key: idx_fk_customer_id
      key_len: 2
          ref: NULL
         rows: 2637
        Extra: Using where
1 row in set (0.00 sec)

(root@localhost) [sakila]> explain select * from payment where customer_id in (200,300,400)\G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: payment
         type: range
possible_keys: idx_fk_customer_id
          key: idx_fk_customer_id
      key_len: 2
          ref: NULL
         rows: 86
        Extra: Using index condition
1 row in set (0.00 sec)

4、ref
非唯一性索引扫描或者，返回匹配某个单独值的所有行。常见于使用非唯一索引即唯一索引的非唯一前缀进行的查找
(root@localhost) [sakila]> explain select * from payment where customer_id=305\G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: payment
         type: ref
possible_keys: idx_fk_customer_id
          key: idx_fk_customer_id
      key_len: 2
          ref: const
         rows: 25
        Extra: 
1 row in set (0.00 sec)

idx_fk_customer_id为表payment上的外键索引，且存在多个不不唯一的值，如下查询
(root@localhost) [sakila]> select customer_id,count(*) from payment group by customer_id
    -> limit 2;
+-------------+----------+
| customer_id | count(*) |
+-------------+----------+
|           1 |       32 |
|           2 |       27 |
+-------------+----------+

-- 下面是非唯一前缀索引使用ref的示例
(root@localhost) [sakila]> create index idx_fisrt_last_name on customer(first_name,last_name);
Query OK, 599 rows affected (0.09 sec)
Records: 599  Duplicates: 0  Warnings: 0

(root@localhost) [sakila]> select first_name,count(*) from customer group by first_name 
    -> having count(*)>1 limit 2;
+------------+----------+
| first_name | count(*) |
+------------+----------+
| JAMIE      |        2 |
| JESSIE     |        2 |
+------------+----------+
2 rows in set (0.00 sec)

(root@localhost) [sakila]> explain select first_name from customer where first_name='JESSIE'\G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: customer
         type: ref
possible_keys: idx_fisrt_last_name
          key: idx_fisrt_last_name
      key_len: 137
          ref: const
         rows: 2
        Extra: Using where; Using index
1 row in set (0.00 sec)

(root@localhost) [sakila]> alter table customer drop index idx_fisrt_last_name;
Query OK, 599 rows affected (0.03 sec)
Records: 599  Duplicates: 0  Warnings: 0

--下面演示出现在join是ref的示例
(root@localhost) [sakila]> explain select b.*,a.* from payment a inner join
    -> customer b on a.customer_id=b.customer_id\G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: b
         type: ALL
possible_keys: PRIMARY
          key: NULL
      key_len: NULL
          ref: NULL
         rows: 599
        Extra: NULL
*************************** 2. row ***************************
           id: 1
  select_type: SIMPLE
        table: a
         type: ref
possible_keys: idx_fk_customer_id
          key: idx_fk_customer_id
      key_len: 2
          ref: sakila.b.customer_id
         rows: 13
        Extra: NULL
2 rows in set (0.01 sec)

5、eq_ref
类似于ref，其差别在于使用的索引为唯一索引，对于每个索引键值，表中只有一条记录与之匹配。
多见于主键扫描或者索引唯一扫描。
(root@localhost) [sakila]> explain select * from film a join film_text b 
    -> on a.film_id=b.film_id;
+----+-------------+-------+--------+---------------+---------+---------+------------------+------+-------------+
| id | select_type | table | type   | possible_keys | key     | key_len | ref              | rows | Extra       |
+----+-------------+-------+--------+---------------+---------+---------+------------------+------+-------------+
|  1 | SIMPLE      | b     | ALL    | PRIMARY       | NULL    | NULL    | NULL             | 1000 | NULL        |
|  1 | SIMPLE      | a     | eq_ref | PRIMARY       | PRIMARY | 2       | sakila.b.film_id |    1 | Using where |
+----+-------------+-------+--------+---------------+---------+---------+------------------+------+-------------+


(root@localhost) [sakila]> explain select title from film where film_id=5;
+----+-------------+-------+-------+---------------+---------+---------+-------+------+-------+
| id | select_type | table | type  | possible_keys | key     | key_len | ref   | rows | Extra |
+----+-------------+-------+-------+---------------+---------+---------+-------+------+-------+
|  1 | SIMPLE      | film  | const | PRIMARY       | PRIMARY | 2       | const |    1 | NULL  |
+----+-------------+-------+-------+---------------+---------+---------+-------+------+-------+

6、const、system：
当MySQL对查询某部分进行优化，这个匹配的行的其他列值可以转换为一个常量来处理。
如将主键或者唯一索引置于where列表中，MySQL就能将该查询转换为一个常量
(root@localhost) [sakila]> create table t1(id int,ename varchar(20) unique);
Query OK, 0 rows affected (0.05 sec)

(root@localhost) [sakila]> insert into t1 values(1,'robin'),(2,'jack'),(3,'henry');
Query OK, 3 rows affected (0.00 sec)
Records: 3  Duplicates: 0  Warnings: 0

(root@localhost) [sakila]> explain select * from (select * from t1 where ename='robin')x;
+----+-------------+------------+--------+---------------+-------+---------+-------+------+-------+
| id | select_type | table      | type   | possible_keys | key   | key_len | ref   | rows | Extra |
+----+-------------+------------+--------+---------------+-------+---------+-------+------+-------+
|  1 | PRIMARY     | <derived2> | system | NULL          | NULL  | NULL    | NULL  |    1 | NULL  |
|  2 | DERIVED     | t1         | const  | ename         | ename | 23      | const |    1 | NULL  |
+----+-------------+------------+--------+---------------+-------+---------+-------+------+-------+
2 rows in set (0.00 sec)

7、type=NULL
MySQL不用访问表或者索引就可以直接得到结果
(root@localhost) [sakila]> explain select sysdate();
+----+-------------+-------+------+---------------+------+---------+------+------+----------------+
| id | select_type | table | type | possible_keys | key  | key_len | ref  | rows | Extra          |
+----+-------------+-------+------+---------------+------+---------+------+------+----------------+
|  1 | SIMPLE      | NULL  | NULL | NULL          | NULL | NULL    | NULL | NULL | No tables used |
+----+-------------+-------+------+---------------+------+---------+------+------+----------------+
1 row in set (0.00 sec)