MySQL之explain详解

前言

开发环境：MySQL5.7.31

参考的博文：青石路大佬的博文https://www.cnblogs.com/youzhibing/p/11909681.html

在日常工作中，我们会有时会开慢查询去记录一些执行时间比较久的SQL语句，找出这些SQL语句并不意味着完事了，些时我们常常用到explain这个命令来查看一个这些SQL语句的执行计划，查看该SQL语句有没有使用上了索引，有没有做全表扫描，这都可以通过explain命令来查看。也就是说explain是MySQL的一个命令，用来查看 SQL 的执行计划（SQL 如何执行），根据其输出结果，我们能够知道以下信息：表的读取顺序，数据读取类型，哪些索引可以使用，哪些索引实际使用了，表之间的连接类型，每张表有多少行被优化器查询等信息，根据这些信息，我们可以找出 SQL 慢的原因，并做针对性的优化。

创建表：

DROP TABLE IF EXISTS tbl_user;
CREATE TABLE tbl_user (
  id INT(11) UNSIGNED NOT NULL AUTO_INCREMENT COMMENT '自增主键',
  user_name VARCHAR(50) NOT NULL COMMENT '用户名',
  sex TINYINT(1) NOT NULL COMMENT '性别, 1:男，0:女',
  create_time datetime NOT NULL COMMENT '创建时间',
  update_time datetime NOT NULL COMMENT '更新时间',
    remark VARCHAR(255) NOT NULL DEFAULT '' COMMENT '备注',
  PRIMARY KEY (id)
) COMMENT='用户表';

DROP TABLE IF EXISTS tbl_user_login_log;
CREATE TABLE tbl_user_login_log (
  id INT(11) UNSIGNED NOT NULL AUTO_INCREMENT COMMENT '自增主键',
  user_name VARCHAR(50) NOT NULL COMMENT '用户名',
  ip VARCHAR(15) NOT NULL COMMENT '登录IP',
  client TINYINT(1) NOT NULL COMMENT '登录端, 1:android, 2:ios, 3:PC, 4:H5',
  create_time datetime NOT NULL COMMENT '创建时间',
  PRIMARY KEY (id)
) COMMENT='登录日志';
INSERT INTO tbl_user(user_name,sex,create_time,update_time,remark) VALUES
('何天香',1,NOW(), NOW(),'朗眉星目，一表人材'),
('薛沉香',0,NOW(), NOW(),'天星楼的总楼主薛摇红的女儿，也是天星楼的少总楼主，体态丰盈，乌发飘逸，指若春葱，袖臂如玉，风姿卓然，高贵典雅，人称“天星绝香”的武林第一大美女'),
('慕容兰娟',0,NOW(), NOW(),'武林东南西北四大世家之北世家慕容长明的独生女儿，生得玲珑剔透，粉雕玉琢，脾气却是刚烈无比，又喜着火红，所以人送绰号“火凤凰”，是除天星楼薛沉香之外的武林第二大美女'),
('苌婷',0,NOW(), NOW(),'当今皇上最宠爱的侄女，北王府的郡主，腰肢纤细，遍体罗绮，眉若墨画，唇点樱红；虽无沉香之雅重，兰娟之热烈，却别现出一种空灵'),
('柳含姻',0,NOW(), NOW(),'武林四绝之一的添愁仙子董婉婉的徒弟，体态窈窕，姿容秀丽，真个是秋水为神玉为骨，芙蓉如面柳如腰，眉若墨画，唇若点樱，不弱西子半分，更胜玉环一筹; 摇红楼、听雨轩，琵琶一曲值千金!'),
('李凝雪',0,NOW(), NOW(),'李相国的女儿，神采奕奕，英姿飒爽，爱憎分明'),
('周遗梦',0,NOW(), NOW(),'音神传人，湘妃竹琴的拥有者，云髻高盘，穿了一身黑色蝉翼纱衫，愈觉得冰肌玉骨，粉面樱唇，格外娇艳动人'),
('叶留痕',0,NOW(), NOW(),'圣域圣女，肤白如雪，白衣飘飘，宛如仙女一般，微笑中带着说不出的柔和之美'),
('郭疏影',0,NOW(), NOW(),'扬灰右使的徒弟，秀发细眉，玉肌丰滑，娇润脱俗'),
('钟钧天',0,NOW(), NOW(),'天界，玄天九部 - 钧天部的部主，超凡脱俗，仙气逼人'),
('王雁云',0,NOW(), NOW(),'尘缘山庄二小姐，刁蛮任性'),
('许侍霜',0,NOW(), NOW(),'药王谷谷主女儿，医术高明'),
('冯黯凝',0,NOW(), NOW(),'桃花门门主，娇艳如火，千娇百媚');
INSERT INTO tbl_user_login_log(user_name, ip, client, create_time) VALUES
('薛沉香', '10.53.56.78',2, '2019-10-12 12:23:45'),
('苌婷', '10.53.56.78',2, '2019-10-12 22:23:45'),
('慕容兰娟', '10.53.56.12',1, '2018-08-12 22:23:45'),
('何天香', '10.53.56.12',1, '2019-10-19 10:23:45'),
('柳含姻', '198.11.132.198',2, '2018-05-12 22:23:45'),
('冯黯凝', '198.11.132.198',2, '2018-11-11 22:23:45'),
('周遗梦', '198.11.132.198',2, '2019-06-18 22:23:45'),
('郭疏影', '220.181.38.148',3, '2019-10-21 09:45:56'),
('薛沉香', '220.181.38.148',3, '2019-10-26 22:23:45'),
('苌婷', '104.69.160.60',4, '2019-10-12 10:23:45'),
('王雁云', '104.69.160.61',4, '2019-10-16 20:23:45'),
('李凝雪', '104.69.160.62',4, '2019-10-17 20:23:45'),
('许侍霜', '104.69.160.63',4, '2019-10-18 20:23:45'),
('叶留痕', '104.69.160.64',4, '2019-10-19 20:23:45'),
('王雁云', '104.69.160.65',4, '2019-10-20 20:23:45'),
('叶留痕', '104.69.160.66',4, '2019-10-21 20:23:45');

试一试：

普通的查询语句

mysql> select * from tbl_user where user_name = "薛沉香";
+----+-----------+-----+---------------------+---------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| id | user_name | sex | create_time         | update_time         | remark                                                                                                                                                                                                                |
+----+-----------+-----+---------------------+---------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|  2 | 薛沉香    |   0 | 2021-02-13 21:28:32 | 2021-02-13 21:28:32 | 天星楼的总楼主薛摇红的女儿，也是天星楼的少总楼主，体态丰盈，乌发飘逸，指若春葱，袖臂如玉，风姿卓然，高贵典雅，人称“天星绝香”的武林第一大美女                                                                          |
+----+-----------+-----+---------------------+---------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

explain只是在查询语句前加上explain即可：

mysql> explain select * from tbl_user where user_name = "薛沉香";
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table    | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tbl_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |    10.00 | Using where |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

我们可以看到这个语句返回来的结果似乎并不存在表中，其实这些字段都是执行计划，通过这些字段我们可以判断这条sql语句是如何执行的，我们该如何进行优化。

主要就是如下字段：

+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table    | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+

接下来逐一看看是什么意思

explain字段解释

id（不重要，了解即可）

选择标识符

输出的是整数，用来标识整个 SQL 的执行顺序。id 如果相同，从上往下依次执行；id不同，id 值越大，执行优先级越高，越先被执行；如果行引用其他行的并集结果，则该值可以为NULL

select_type（一般重要，学会几个类型用于面试加分）

就是select类型，表示查询的类型。

有以下多种：

(1) SIMPLE(简单SELECT，不使用UNION或子查询等)

(2) PRIMARY(子查询中最外层查询，查询中若包含任何复杂的子部分，最外层的select被标记为PRIMARY)

(3) UNION(UNION中的第二个或后面的SELECT语句)

(4) DEPENDENT UNION(UNION中的第二个或后面的SELECT语句，取决于外面的查询)

(5) UNION RESULT(UNION的结果，union语句中第二个select开始后面所有select)

(6) SUBQUERY(子查询中的第一个SELECT，结果不依赖于外部查询)

(7) DEPENDENT SUBQUERY(子查询中的第一个SELECT，依赖于外部查询)

(8) DERIVED(派生表的SELECT, FROM子句的子查询)

(9) UNCACHEABLE SUBQUERY(一个子查询的结果不能被缓存，必须重新评估外链接的第一行)

SIMPLLE、PRIMARY、SUBQUERY、DERIVED 这 4 个在实际工作中碰到的会比较多，记住这四个含义就行了，其他了解即可

SIMPLE

SIMPLE：简单的 SELECT 查询，没有 UNION 或者子查询，包括单表查询或者多表 JOIN 查询

示例：

EXPLAIN 
SELECT t1.*, tu.sex, tu.remark 
FROM tbl_user_login_log t1 
LEFT JOIN tbl_user tu 
ON t1.user_name = tu.user_name;

结果：

+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------------------------------------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra                                              |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------------------------------------------+
|  1 | SIMPLE      | t1    | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   16 |   100.00 | NULL                                               |
|  1 | SIMPLE      | tu    | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |   100.00 | Using where; Using join buffer (Block Nested Loop) |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------------------------------------------+
2 rows in set, 1 warning (0.00 sec)

说明：

select_type列都是SIMPLE，说明是简单的select语句，不使用UNION或子查询等
补充一下id的讲解，这里id都是1，执行就按照从上到下的顺序执行，也就是table列，想执行tl再执行tu

PRIMARY

PRIMARY：最外层的 select 查询，常见于子查询或 UNION 查询，最外层的查询被标识为 PRIMARY

示例：

EXPLAIN 
SELECT * FROM tbl_user tu 
WHERE EXISTS(
  SELECT * FROM tbl_user_login_log tull
  WHERE tull.`user_name` = tu.`user_name`
);

结果：

+----+--------------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type        | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+--------------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | PRIMARY            | tu    | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |   100.00 | Using where |
|  2 | DEPENDENT SUBQUERY | tull  | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   16 |    10.00 | Using where |
+----+--------------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
2 rows in set, 2 warnings (0.00 sec)

说明：

注意id为1的，它是最外层的select，所以是最慢执行的，所以id最小
理解primary是最外层select即可

UNION

UNION：UNION 操作的第二个或之后的 SELECT，不依赖于外部查询的结果集（外部查询指的就是 PRIMARY 对应的 SELECT）

示例：

EXPLAIN 
SELECT * FROM tbl_user 
WHERE user_name = '薛沉香'
UNION
SELECT * FROM tbl_user WHERE user_name = '慕容兰娟';

结果：

+----+--------------+------------+------------+------+---------------+------+---------+------+------+----------+-----------------+
| id | select_type  | table      | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra           |
+----+--------------+------------+------------+------+---------------+------+---------+------+------+----------+-----------------+
|  1 | PRIMARY      | tbl_user   | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |    10.00 | Using where     |
|  2 | UNION        | tbl_user   | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |    10.00 | Using where     |
| NULL | UNION RESULT | <union1,2> | NULL       | ALL  | NULL          | NULL | NULL    | NULL | NULL |     NULL | Using temporary |
+----+--------------+------------+------------+------+---------------+------+---------+------+------+----------+-----------------+
3 rows in set, 1 warning (0.00 sec)

说明：

简单来说就是若第二个SELECT出现在UNION之后，则被标记为UNION

DEPENDENT UNION

DEPENDENT UNION：UNION 操作的第二个或之后的 SELECT，依赖于外部查询的结果集

示例：

EXPLAIN 
SELECT * FROM tbl_user_login_log WHERE user_name IN(
  SELECT user_name FROM tbl_user WHERE id = 2
  UNION
  SELECT user_name FROM tbl_user WHERE id = 3
);

结果：

原sql语句的结果：

+----+--------------+----------------+--------+---------------------+
| id | user_name    | ip             | client | create_time         |
+----+--------------+----------------+--------+---------------------+
|  1 | 薛沉香       | 10.53.56.78    |      2 | 2019-10-12 12:23:45 |
|  3 | 慕容兰娟     | 10.53.56.12    |      1 | 2018-08-12 22:23:45 |
|  9 | 薛沉香       | 220.181.38.148 |      3 | 2019-10-26 22:23:45 |
+----+--------------+----------------+--------+---------------------+
3 rows in set (0.00 sec)

explain执行计划结果：

+----+--------------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-----------------+
| id | select_type        | table              | partitions | type  | possible_keys | key     | key_len | ref   | rows | filtered | Extra           |
+----+--------------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-----------------+
|  1 | PRIMARY            | tbl_user_login_log | NULL       | ALL   | NULL          | NULL    | NULL    | NULL  |   16 |   100.00 | Using where     |
|  2 | DEPENDENT SUBQUERY | tbl_user           | NULL       | const | PRIMARY       | PRIMARY | 4       | const |    1 |   100.00 | NULL            |
|  3 | DEPENDENT UNION    | tbl_user           | NULL       | const | PRIMARY       | PRIMARY | 4       | const |    1 |   100.00 | NULL            |
| NULL | UNION RESULT       | <union2,3>         | NULL       | ALL   | NULL          | NULL    | NULL    | NULL  | NULL |     NULL | Using temporary |
+----+--------------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-----------------+
4 rows in set, 1 warning (0.00 sec)

说明：

很明显primary就是指最外层的select，它的优先级最后，id最小
id = 3 为DEPENDENT UNION，它是union中第二个查询
依赖于外部查询的结果集就是说取决于外面的查询的意思

UNION RESULT

UNION RESULT：UNION 的结果（如果是 UNION ALL 则无此结果）

说明：

UNION和DEPENDENT UNION的例子中我们都能在结果看到UNION RESULT，表示从UNION表获取结果的SELECT

SUBQUERY

SUBQUERY：子查询中的第一个 SELECT 查询，不依赖于外部查询的结果集

示例：

EXPLAIN 
SELECT * FROM tbl_user_login_log WHERE user_name = (
  SELECT user_name FROM tbl_user WHERE id = 2
);

结果：

+----+-------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-------------+
| id | select_type | table              | partitions | type  | possible_keys | key     | key_len | ref   | rows | filtered | Extra       |
+----+-------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-------------+
|  1 | PRIMARY     | tbl_user_login_log | NULL       | ALL   | NULL          | NULL    | NULL    | NULL  |   16 |    10.00 | Using where |
|  2 | SUBQUERY    | tbl_user           | NULL       | const | PRIMARY       | PRIMARY | 4       | const |    1 |   100.00 | NULL        |
+----+-------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-------------+
2 rows in set, 1 warning (0.00 sec)

说明：

没啥好说的，看table和id就知道SUBQUERY指代的执行顺序

DEPENDENT SUBQUERY

DEPENDENT SUBQUERY：子查询中的第一个select查询，依赖于外部查询的结果集

示例：

EXPLAIN 
SELECT * FROM tbl_user_login_log WHERE user_name IN (
  SELECT user_name FROM tbl_user WHERE id = 2
  UNION ALL
  SELECT user_name FROM tbl_user WHERE id = 3
);

结果：

+----+--------------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-------------+
| id | select_type        | table              | partitions | type  | possible_keys | key     | key_len | ref   | rows | filtered | Extra       |
+----+--------------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-------------+
|  1 | PRIMARY            | tbl_user_login_log | NULL       | ALL   | NULL          | NULL    | NULL    | NULL  |   16 |   100.00 | Using where |
|  2 | DEPENDENT SUBQUERY | tbl_user           | NULL       | const | PRIMARY       | PRIMARY | 4       | const |    1 |   100.00 | NULL        |
|  3 | DEPENDENT UNION    | tbl_user           | NULL       | const | PRIMARY       | PRIMARY | 4       | const |    1 |   100.00 | NULL        |
+----+--------------------+--------------------+------------+-------+---------------+---------+---------+-------+------+----------+-------------+
3 rows in set, 1 warning (0.00 sec)

说明：

DEPENDENT UNION的例子中我们也可以见到DEPENDENT SUBQUERY
就是union中第一个select，因为外部还有select，所以带有个DEPENDENT

DERIVED

DERIVED：派生表（临时表），常见于 FROM 子句中有子查询的情况

示例：

EXPLAIN 
SELECT * FROM (
  SELECT * FROM tbl_user WHERE user_name = '薛沉香'
) t;

结果：

+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table    | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tbl_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |    10.00 | Using where |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

说明：

你一定会奇怪为什么出现的是SIMPLE而不是DERIVED，这是因为MySQL5.7 中对 Derived table 做了一个新特性，该特性允许将符合条件的 Derived table 中的子表与父查询的表合并进行直接JOIN，从而简化了执行计划，同时也提高了执行效率；默认情况下，MySQL5.7 中这个特性是开启的
可通过 SET SESSION optimizer_switch='derived_merge=on|off' 来开启或关闭当前 SESSION 的该特性。

mysql> SET SESSION optimizer_switch='derived_merge=off';
Query OK, 0 rows affected (0.00 sec)

mysql> EXPLAIN
-> SELECT * FROM (
-> SELECT * FROM tbl_user WHERE user_name = '薛沉香'
-> ) t;
+----+-------------+------------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+------------+------------+------+---------------+------+---------+------+------+----------+-------------+
| 1 | PRIMARY | | NULL | ALL | NULL | NULL | NULL | NULL | 2 | 100.00 | NULL |
| 2 | DERIVED | tbl_user | NULL | ALL | NULL | NULL | NULL | NULL | 13 | 10.00 | Using where |
+----+-------------+------------+------------+------+---------------+------+---------+------+------+----------+-------------+
2 rows in set, 1 warning (0.00 sec)

table（一般都懂）

显示该行正在访问哪个表

partitions（不重要）

查询进行匹配的分区，对于非分区表，该值为NULL。

type（很重要）

这列很重要,显示了连接使用了哪种类别,有无使用索引。下面从性能最好到性能最差排序解析：

system

该表只有一行（=系统表），是 const 类型的特例

const

确定只有一行匹配的时候，mysql 优化器会在查询前读取它并且只读取一次，速度非常快。用于 primary key 或 unique 索引中有常亮值比较的情形

mysql> SET SESSION optimizer_switch='derived_merge=off';
Query OK, 0 rows affected (0.00 sec)

mysql> EXPLAIN SELECT * FROM (SELECT * FROM tbl_user WHERE id = 2) a;
+----+-------------+------------+------------+--------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table      | partitions | type   | possible_keys | key     | key_len | ref   | rows | filtered | Extra |
+----+-------------+------------+------------+--------+---------------+---------+---------+-------+------+----------+-------+
|  1 | PRIMARY     | <derived2> | NULL       | system | NULL          | NULL    | NULL    | NULL  |    1 |   100.00 | NULL  |
|  2 | DERIVED     | tbl_user   | NULL       | const  | PRIMARY       | PRIMARY | 4       | const |    1 |   100.00 | NULL  |
+----+-------------+------------+------------+--------+---------------+---------+---------+-------+------+----------+-------+
2 rows in set, 1 warning (0.00 sec)

eq_ref

对于每个来自于前面的表的行，从该表最多只返回一条符合条件的记录。当连接使用的索引是 PRIMARY KEY 或 UNIQUE NOT NULL 索引时使用，非常高效

mysql> ALTER TABLE tbl_user ADD UNIQUE (user_name);
Query OK, 0 rows affected (0.02 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> EXPLAIN SELECT * FROM tbl_user_login_log tl
    -> LEFT JOIN tbl_user tu ON tl.user_name = tu.user_name;
+----+-------------+-------+------------+--------+---------------+-----------+---------+-------------------+------+----------+-------+
| id | select_type | table | partitions | type   | possible_keys | key       | key_len | ref               | rows | filtered | Extra |
+----+-------------+-------+------------+--------+---------------+-----------+---------+-------------------+------+----------+-------+
|  1 | SIMPLE      | tl    | NULL       | ALL    | NULL          | NULL      | NULL    | NULL              |   16 |   100.00 | NULL  |
|  1 | SIMPLE      | tu    | NULL       | eq_ref | user_name     | user_name | 152     | test.tl.user_name |    1 |   100.00 | NULL  |
+----+-------------+-------+------------+--------+---------------+-----------+---------+-------------------+------+----------+-------+
2 rows in set, 1 warning (0.00 sec)

ref

索引访问，也称索引查找，它返回所有匹配某个单个值的行。此类型通常出现在多表的 JOIN 查询, 针对于非 UNIQUE 或非 PRIMARY KEY, 或者是使用了最左前缀规则索引的查询，换句话说，如果 JOIN 不能基于关键字选择单个行的话，则使用ref

mysql> ALTER TABLE tbl_user_login_log ADD INDEX idx_user_name (user_name);
Query OK, 0 rows affected (0.03 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> EXPLAIN SELECT * FROM tbl_user_login_log tl
    -> RIGHT JOIN tbl_user tu ON tl.user_name = tu.user_name;
+----+-------------+-------+------------+------+---------------+---------------+---------+-------------------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key           | key_len | ref               | rows | filtered | Extra |
+----+-------------+-------+------------+------+---------------+---------------+---------+-------------------+------+----------+-------+
|  1 | SIMPLE      | tu    | NULL       | ALL  | NULL          | NULL          | NULL    | NULL              |   13 |   100.00 | NULL  |
|  1 | SIMPLE      | tl    | NULL       | ref  | idx_user_name | idx_user_name | 152     | test.tu.user_name |    1 |   100.00 | NULL  |
+----+-------------+-------+------------+------+---------------+---------------+---------+-------------------+------+----------+-------+
2 rows in set, 1 warning (0.00 sec)

fulltext

当使用全文索引时会用到，这种索引一般用不到，会用专门的搜索服务(solr、elasticsearch等)来替代

ref_or_null

类似ref，但是添加了可以专门搜索 NULL 的行

ALTER TABLE tbl_user ADD COLUMN weapon VARCHAR(50) COMMENT '武器';
UPDATE tbl_user SET weapon = '碧玉笛' WHERE user_name = '何天香';
UPDATE tbl_user SET weapon = '天问剑' WHERE user_name = '薛沉香';
UPDATE tbl_user SET weapon = '流星剑' WHERE user_name = '慕容兰娟';
UPDATE tbl_user SET weapon = '相思刀' WHERE user_name = '郭疏影';
UPDATE tbl_user SET weapon = '湘妃竹琴' WHERE user_name = '周遗梦';
UPDATE tbl_user SET weapon = '玉琵琶' WHERE user_name = '柳含姻';
ALTER TABLE tbl_user ADD INDEX idx_weapon (weapon);
EXPLAIN SELECT * FROM tbl_user WHERE weapon = '湘妃竹琴' OR weapon IS NULL;

+----+-------------+----------+------------+-------------+---------------+------------+---------+-------+------+----------+-----------------------+
| id | select_type | table    | partitions | type        | possible_keys | key        | key_len | ref   | rows | filtered | Extra                 |
+----+-------------+----------+------------+-------------+---------------+------------+---------+-------+------+----------+-----------------------+
|  1 | SIMPLE      | tbl_user | NULL       | ref_or_null | idx_weapon    | idx_weapon | 153     | const |    8 |   100.00 | Using index condition |
+----+-------------+----------+------------+-------------+---------------+------------+---------+-------+------+----------+-----------------------+
1 row in set, 1 warning (0.00 sec)

这个是有前提条件的，前提为 weapon 列有索引，且 weapon 列存在 NULL

index_merge

该访问类型使用了索引合并优化方法

mysql> EXPLAIN SELECT * FROM tbl_user WHERE id = 1 OR weapon = '流星剑';
+----+-------------+----------+------------+-------------+--------------------+--------------------+---------+------+------+----------+----------------------------------------------+
| id | select_type | table    | partitions | type        | possible_keys      | key                | key_len | ref  | rows | filtered | Extra                                        |
+----+-------------+----------+------------+-------------+--------------------+--------------------+---------+------+------+----------+----------------------------------------------+
|  1 | SIMPLE      | tbl_user | NULL       | index_merge | PRIMARY,idx_weapon | PRIMARY,idx_weapon | 4,153   | NULL |    2 |   100.00 | Using union(PRIMARY,idx_weapon); Using where |
+----+-------------+----------+------------+-------------+--------------------+--------------------+---------+------+------+----------+----------------------------------------------+
1 row in set, 1 warning (0.01 sec)

这个同样也是有条件的， id 列和 weapon 列都有单列索引。如果出现 index_merge，并且这类 SQL 后期使用较频繁，可以考虑把单列索引换为组合索引，这样效率更高

unique_subquery

类似于两表连接中被驱动表的 eq_ref 访问方式，unique_subquery 是针对在一些包含 IN 子查询的查询语句中，如果查询优化器决定将 IN 子查询转换为 EXISTS 子查询，而且子查询可以使用到主键或者唯一索引进行等值匹配时，则会使用 unique_subquery

mysql> EXPLAIN SELECT * FROM tbl_user t1 WHERE user_name IN (
    -> SELECT user_name FROM tbl_user t2 WHERE t1.weapon = t2.weapon
    -> ) OR weapon = '相思刀';
+----+--------------------+-------+------------+-----------------+----------------------+-----------+---------+------+------+----------+-------------+
| id | select_type        | table | partitions | type            | possible_keys        | key       | key_len | ref  | rows | filtered | Extra       |
+----+--------------------+-------+------------+-----------------+----------------------+-----------+---------+------+------+----------+-------------+
|  1 | PRIMARY            | t1    | NULL       | ALL             | idx_weapon           | NULL      | NULL    | NULL |   13 |   100.00 | Using where |
|  2 | DEPENDENT SUBQUERY | t2    | NULL       | unique_subquery | user_name,idx_weapon | user_name | 152     | func |    1 |    10.00 | Using where |
+----+--------------------+-------+------------+-----------------+----------------------+-----------+---------+------+------+----------+-------------+
2 rows in set, 2 warnings (0.00 sec)

index_subquery

index_subquery 与 unique_subquery类似，只不过访问子查询中的表时使用的是普通的索引

mysql> EXPLAIN SELECT * FROM tbl_user t1 WHERE weapon IN (
    -> SELECT weapon FROM tbl_user t2 WHERE t1.weapon = '玉琵琶'
    -> ) OR weapon = '相思刀';
+----+--------------------+-------+------------+----------------+---------------+------------+---------+------+------+----------+--------------------------+
| id | select_type        | table | partitions | type           | possible_keys | key        | key_len | ref  | rows | filtered | Extra                    |
+----+--------------------+-------+------------+----------------+---------------+------------+---------+------+------+----------+--------------------------+
|  1 | PRIMARY            | t1    | NULL       | ALL            | idx_weapon    | NULL       | NULL    | NULL |   13 |   100.00 | Using where              |
|  2 | DEPENDENT SUBQUERY | t2    | NULL       | index_subquery | idx_weapon    | idx_weapon | 153     | func |    1 |   100.00 | Using where; Using index |
+----+--------------------+-------+------------+----------------+---------------+------------+---------+------+------+----------+--------------------------+
2 rows in set, 2 warnings (0.00 sec)

range

使用索引来检索给定范围的行，当使用 =、<>、>、>=、<、<=、IS NULL、<=>、BETWEEN 或者 IN 操作符，用常量比较关键字列时，则会使用 rang

mysql> EXPLAIN SELECT * FROM tbl_user WHERE id < 6;
+----+-------------+----------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
| id | select_type | table    | partitions | type  | possible_keys | key     | key_len | ref  | rows | filtered | Extra       |
+----+-------------+----------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tbl_user | NULL       | range | PRIMARY       | PRIMARY | 4       | NULL |    5 |   100.00 | Using where |
+----+-------------+----------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

前提是必须基于索引，也就是 id 上必须有索引

index

当我们可以使用索引覆盖，但需要扫描全部的索引记录时，则会使用 index；进行统计时非常常见

mysql> EXPLAIN SELECT COUNT(*) FROM tbl_user;
+----+-------------+----------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| id | select_type | table    | partitions | type  | possible_keys | key       | key_len | ref  | rows | filtered | Extra       |
+----+-------------+----------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tbl_user | NULL       | index | NULL          | user_name | 152     | NULL |   13 |   100.00 | Using index |
+----+-------------+----------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

all

全表扫描

mysql> EXPLAIN SELECT * FROM tbl_user;
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------+
| id | select_type | table    | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------+
|  1 | SIMPLE      | tbl_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |   100.00 | NULL  |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

possible_keys（了解即可）

展示在这个 SQL 中，可能用到的索引有哪些，但不一定在查询时使用。若为空则表示没有可以使用的索引，此时可以通过检查 WHERE 语句看是否可以引用某些列或者新建索引来提高性能

key（重要）

展示这个 SQL 实际使用的索引，如果没有选择索引，则此列为null，要想强制 MySQL 使用或忽视 possible_keys 列中的索引，在查询中使用 FORCE INDEX、USE INDEX 或者I GNORE INDEX

key_len（了解即可）

展示 MySQL 决定使用的索引长度（字节数）。如果 key 是 NULL，则长度为 NULL

注意：

key_len显示的值为索引字段的最大可能长度，并非实际使用长度，即key_len是根据表定义计算而得，不是通过表内检索出的。
在不损失精确性的情况下，长度越短越好

ref（了解即可）

显示索引的那一列被使用了

rows（了解即可）

根据表统计信息及索引选用情况，大致估算出找到所需的记录所需要读取的行数，也就是说，用的越少越好

filtered（不重要）

展示的是被条件过滤的行数所占需要读到的行（rows 的值）的比例（省略了百分号，显示的是百分号前的值）

extra（很重要）

表示不在其他列但也很重要的额外信息。取值有很多，以下是一些常见的

using index（看见了就安心了）

表示 SQL 使用了使用覆盖索引，而不用回表去查询数据，性能非常不错

mysql> EXPLAIN SELECT user_name FROM tbl_user;
+----+-------------+----------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| id | select_type | table    | partitions | type  | possible_keys | key       | key_len | ref  | rows | filtered | Extra       |
+----+-------------+----------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tbl_user | NULL       | index | NULL          | user_name | 152     | NULL |   13 |   100.00 | Using index |
+----+-------------+----------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

using where

表示存储引擎搜到记录后进行了后过滤(POST-FILTER)，如果查询未能使用索引，using where 的作用只是提醒我们 mysql 要用 where 条件过滤结果集

mysql> EXPLAIN SELECT * FROM tbl_user WHERE sex = 0;
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table    | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tbl_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |    10.00 | Using where |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

using temporary

表示 mysql 需要使用临时表来存储结果集，常见于排序和分组查询

mysql> EXPLAIN SELECT sex, count(*) FROM tbl_user GROUP BY sex;
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+---------------------------------+
| id | select_type | table    | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra                           |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+---------------------------------+
|  1 | SIMPLE      | tbl_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |   100.00 | Using temporary; Using filesort |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+---------------------------------+
1 row in set, 1 warning (0.00 sec)

using filesort

表示 mysql 无法利用索引直接完成排序（排序的字段不是索引字段），此时会用到缓冲空间（内存或者磁盘）来进行排序；一般出现该值，则表示 SQL 要进行优化了，它对 CPU 的消耗是比较大的

mysql> EXPLAIN SELECT * FROM tbl_user ORDER BY sex DESC;
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+----------------+
| id | select_type | table    | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra          |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+----------------+
|  1 | SIMPLE      | tbl_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   13 |   100.00 | Using filesort |
+----+-------------+----------+------------+------+---------------+------+---------+------+------+----------+----------------+
1 row in set, 1 warning (0.00 sec)

impossible where

查询语句的WHERE子句永远为 FALSE 时将会提示该额外信息

mysql> EXPLAIN SELECT * FROM tbl_user WHERE id IS NULL;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+------------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra            |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+------------------+
|  1 | SIMPLE      | NULL  | NULL       | NULL | NULL          | NULL | NULL    | NULL | NULL |     NULL | Impossible WHERE |
+----+-------------+-------+---------

总结

EXPLAIN不会告诉你关于触发器、存储过程的信息或用户自定义函数对查询的影响情况
EXPLAIN不考虑各种Cache
EXPLAIN不能显示MySQL在执行查询时所作的优化工作
部分统计信息是估算的，并非精确值
EXPALIN只能解释SELECT操作，其他操作要重写为SELECT后查看执行计划。

好东西，explain思维导图：https://www.processon.com/mindmap/6027f4f5f346fb64f565cf90

相关阅读:
mvc control 请求两次问题
 Jquery跨域获得Json
使用ajax跨域withCredentials的作用
 以下是关于ASP.NET中保存各种信息的对象的比较，理解这些对象的原理，对制作完善的程序来说是相当有必要的(摘至互联网,并非原创--xukunping)
【算法】字符串近似搜索（转）
字符串相似度算法（编辑距离算法 Levenshtein Distance）（转）
图像相似度算法的C#实现及测评
 求时间差的sql语句。比如如下数据
 微信公众平台开发之微信access_token如何有效长期保存
 PowerDesigner打开设计文件后提示failed to read the fileXXX的解决办法
原文地址：https://www.cnblogs.com/kylinxxx/p/14403135.html