一文學會MySQL的explain工具
阿新 • • 發佈：2020-08-03
## 開篇說明
(1) 本文將細緻介紹MySQL的explain工具，是下一篇《一文讀懂MySQL的索引機制及查詢優化》的準備篇。
(2) 本文主要基於MySQL`5.7`版本(`https://dev.mysql.com/doc/refman/5.7/en/`)，MySQL`8.x`版本可另行翻閱對應版本文件(`https://dev.mysql.com/doc/refman/8.0/en/`)。
(3) 演示過程中的建庫、建表、建索引等語句僅為了測試explain工具的使用，並未考慮實際應用場景的合理性。
## explain工具介紹
相關文件：
`https://dev.mysql.com/doc/refman/5.7/en/explain.html`
`https://dev.mysql.com/doc/refman/5.7/en/using-explain.html`
> EXPLAIN is used to obtain a query execution plan (that is, an explanation of how MySQL would execute a query).
簡單翻譯一下，就是explain用於獲取查詢執行計劃（即MySQL是如何執行一個查詢的）。
工作中，我們會遇到慢查詢，這個時候我們就可以在`select`語句之前增加`explain`關鍵字，模擬MySQL優化器執行SQL語句，從而分析該SQL語句有沒有用上索引、是否全表掃描、能否進一步優化等。
還是來個快速入門的案例比較直觀，依次在mysql的命令列執行下面幾條語句(建庫、建表sql指令碼見下面的`資料準備`部分)：
```sh
mysql> use `explain_test`;
mysql> select * from tb_hero where hero_name = '李尋歡' and book_id = 1;
mysql> explain select * from tb_hero where hero_name = '李尋歡' and book_id = 1;
mysql> show warnings \G
```
得到下面的輸出：
```sh
mysql> use `explain_test`;
Database changed
mysql> select * from tb_hero where hero_name = '李尋歡' and book_id = 1;
+---------+-----------+--------------+---------+
| hero_id | hero_name | skill | book_id |
+---------+-----------+--------------+---------+
| 1 | 李尋歡 | 小李飛刀 | 1 |
+---------+-----------+--------------+---------+
1 row in set (0.00 sec)
mysql> explain select * from tb_hero where hero_name = '李尋歡' and book_id = 1;
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-------+
| 1 | SIMPLE | tb_hero | NULL | ref | idx_book_id_hero_name | idx_book_id_hero_name | 136 | const,const | 1 | 100.00 | NULL |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
mysql> show warnings \G
*************************** 1. row ***************************
Level: Note
Code: 1003
Message: /* select#1 */ select `explain_test`.`tb_hero`.`hero_id` AS `hero_id`,`explain_test`.`tb_hero`.`hero_name` AS `hero_name`,`explain_test`.`tb_hero`.`skill` AS `skill`,`explain_test`.`tb_hero`.`book_id` AS `book_id` from `explain_test`.`tb_hero` where ((`explain_test`.`tb_hero`.`book_id` = 1) and (`explain_test`.`tb_hero`.`hero_name` = '李尋歡'))
1 row in set (0.00 sec)
```
先別急`explain`語句輸出結果每一列表示什麼意思(後面會具體描述)，用`show warnings`命令可以得到優化後的查詢語句大致長什麼樣子。
補充：
- 有關`show warnings`更詳細的使用見`https://dev.mysql.com/doc/refman/5.7/en/show-warnings.html`
- 有關獲取`explain`額外的輸出資訊見`https://dev.mysql.com/doc/refman/5.7/en/explain-extended.html`
原SQL語句：
```sh
select * from tb_hero where hero_name = '李尋歡' and book_id = 1;
```
優化後的SQL語句：
```sh
select `explain_test`.`tb_hero`.`hero_id` AS `hero_id`,
`explain_test`.`tb_hero`.`hero_name` AS `hero_name`,
`explain_test`.`tb_hero`.`skill` AS `skill`,
`explain_test`.`tb_hero`.`book_id` AS `book_id`
from `explain_test`.`tb_hero`
where ((`explain_test`.`tb_hero`.`book_id` = 1) and (`explain_test`.`tb_hero`.`hero_name` = '李尋歡'))
```
可以看出，MySQL優化器把`*`優化成具體的列名，另外把我`where`中的兩個過濾條件`hero_name`、`book_id`先後順序調換了一下，這種順序調換是概率性事件還是另有文章？
(哈哈哈，(●´ω｀●)留個懸念，本篇僅介紹explain工具，讀了下篇《一文讀懂MySQL的索引機制及查詢優化》後自然豁然開朗)
## 資料準備
為了方便演示explain工具的使用以及輸出結果的含義，準備了一些測試資料，初始化sql指令碼如下:
```sh
-- ----------------------------
-- create database
-- ----------------------------
DROP database IF EXISTS `explain_test`;
create database `explain_test` DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;
-- switch database
use `explain_test`;
-- ----------------------------
-- table structure for `tb_book`
-- ----------------------------
DROP TABLE IF EXISTS `tb_book`;
CREATE TABLE `tb_book` (
`book_id` int(11) NOT NULL,
`book_name` varchar(64) DEFAULT NULL,
`author` varchar(32) DEFAULT NULL,
PRIMARY KEY (`book_id`),
UNIQUE KEY `uk_book_name` (`book_name`) USING BTREE,
INDEX `idx_author` (`author`) USING BTREE
);
BEGIN;
INSERT INTO `tb_book`(`book_id`, `book_name`, `author`) VALUES (1, '多情劍客無情劍', '古龍');
INSERT INTO `tb_book`(`book_id`, `book_name`, `author`) VALUES (2, '笑傲江湖', '金庸');
INSERT INTO `tb_book`(`book_id`, `book_name`, `author`) VALUES (3, '倚天屠龍記', '金庸');
INSERT INTO `tb_book`(`book_id`, `book_name`, `author`) VALUES (4, '射鵰英雄傳', '金庸');
INSERT INTO `tb_book`(`book_id`, `book_name`, `author`) VALUES (5, '絕代雙驕', '古龍');
COMMIT;
-- ----------------------------
-- table structure for `tb_hero`
-- ----------------------------
DROP TABLE IF EXISTS `tb_hero`;
CREATE TABLE `tb_hero` (
`hero_id` int(11) NOT NULL,
`hero_name` varchar(32) DEFAULT NULL,
`skill` varchar(64) DEFAULT NULL,
`book_id` int(11) DEFAULT NULL,
PRIMARY KEY (`hero_id`),
INDEX `idx_book_id_hero_name`(`book_id`, `hero_name`) USING BTREE
);
BEGIN;
INSERT INTO `tb_hero`(`hero_id`, `hero_name`, `skill`, `book_id`) VALUES (1, '李尋歡', '小李飛刀', 1);
INSERT INTO `tb_hero`(`hero_id`, `hero_name`, `skill`, `book_id`) VALUES (2, '令狐沖', '獨孤九劍', 2);
INSERT INTO `tb_hero`(`hero_id`, `hero_name`, `skill`, `book_id`) VALUES (3, '張無忌', '九陽神功', 3);
INSERT INTO `tb_hero`(`hero_id`, `hero_name`, `skill`, `book_id`) VALUES (4, '郭靖', '降龍十八掌', 4);
INSERT INTO `tb_hero`(`hero_id`, `hero_name`, `skill`, `book_id`) VALUES (5, '花無缺', '移花接玉', 5);
INSERT INTO `tb_hero`(`hero_id`, `hero_name`, `skill`, `book_id`) VALUES (6, '任我行', '吸星大法', 2);
COMMIT;
-- ----------------------------
-- Table structure for `tb_book_hero`
-- ----------------------------
DROP TABLE IF EXISTS `tb_book_hero`;
CREATE TABLE `tb_book_hero` (
`book_id` int(11) NOT NULL,
`hero_id` int(11) NOT NULL,
`user_comment` varchar(255) DEFAULT NULL,
PRIMARY KEY (`book_id`, `hero_id`) USING BTREE
) ENGINE = InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
BEGIN;
INSERT INTO `tb_book_hero`(`book_id`, `hero_id`, `user_comment`) VALUES (1, 1, '小李飛刀，例無虛發，奪魂索命，彈指之間');
INSERT INTO `tb_book_hero`(`book_id`, `hero_id`, `user_comment`) VALUES (2, 2, '令狐少俠留步!');
INSERT INTO `tb_book_hero`(`book_id`, `hero_id`, `user_comment`) VALUES (3, 3, '嚐遍世間善惡，歸來仍是少年');
INSERT INTO `tb_book_hero`(`book_id`, `hero_id`, `user_comment`) VALUES (4, 4, '我只要我的靖哥哥!');
INSERT INTO `tb_book_hero`(`book_id`, `hero_id`, `user_comment`) VALUES (5, 5, '風采儒雅亦坦蕩，武藝精深兼明智。');
INSERT INTO `tb_book_hero`(`book_id`, `hero_id`, `user_comment`) VALUES (2, 6, '有人就有恩怨，有恩怨就有江湖，人心即是江湖，你如何退出！');
COMMIT;
```
## explain的輸出結果
相關文件：
`https://dev.mysql.com/doc/refman/5.7/en/explain-output.html`
看一下官方文件顯示的關於explain輸出結果列(`explain output columns`)的含義：
|Column |JSON Name| Meaning|
|--|--|--|
|id| select_id |The SELECT identifier
|select_type| None| The SELECT type|
|table| table_name| The table for the output row|
|partitions| partitions| The matching partitions|
|type| access_type| The join type|
|possible_keys| possible_keys| The possible indexes to choose|
|key| key| The index actually chosen|
|key_len| key_length| The length of the chosen key|
|ref| ref| The columns compared to the index|
|rows| rows| Estimate of rows to be examined|
|filtered| filtered| Percentage of rows filtered by table condition|
|Extra| None| Additional information|
其中`JSON Name`指的是當設定`FORMAT=JSON`時，列名在json中顯示的name，見下面的演示就明白了
```sh
mysql> explain select * from tb_book \G
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: tb_book
partitions: NULL
type: ALL
possible_keys: NULL
key: NULL
key_len: NULL
ref: NULL
rows: 5
filtered: 100.00
Extra: NULL
1 row in set, 1 warning (0.00 sec)
mysql> explain FORMAT=JSON select * from tb_book \G
*************************** 1. row ***************************
EXPLAIN: {
"query_block": {
"select_id": 1,
"cost_info": {
"query_cost": "2.00"
},
"table": {
"table_name": "tb_book",
"access_type": "ALL",
"rows_examined_per_scan": 5,
"rows_produced_per_join": 5,
"filtered": "100.00",
"cost_info": {
"read_cost": "1.00",
"eval_cost": "1.00",
"prefix_cost": "2.00",
"data_read_per_join": "1K"
},
"used_columns": [
"book_id",
"book_name",
"author"
]
}
}
}
1 row in set, 1 warning (0.00 sec)
```
下面重點看一下比較重要的幾個欄位。
### id列
`id`是`select`的唯一標識，有幾個`select`就有幾個id，並且id的順序是按`select`出現的順序增長的，id值越大執行優先順序越高，id相同則從上往下執行，id為NULL最後執行。
為了驗證上面的結論，臨時關閉mysql`5.7`對子查詢(`sub queries`)產生的衍生表(`derived tables`)的合併優化
```sh
set session optimizer_switch='derived_merge=off';
```
詳情見：
`https://dev.mysql.com/doc/refman/5.7/en/switchable-optimizations.html`
`https://dev.mysql.com/doc/refman/5.7/en/derived-table-optimization.html`
```sh
mysql> set session optimizer_switch='derived_merge=off';
Query OK, 0 rows affected (0.00 sec)
mysql> select (select count(1) from tb_book) as book_count, (select count(1) from tb_hero) as hero_count from (select * from tb_book_hero) as book_hero;
+------------+------------+
| book_count | hero_count |
+------------+------------+
| 5 | 6 |
| 5 | 6 |
| 5 | 6 |
| 5 | 6 |
| 5 | 6 |
| 5 | 6 |
+------------+------------+
6 rows in set (0.00 sec)
mysql> explain select (select count(1) from tb_book) as book_count, (select count(1) from tb_hero) as hero_count from (select * from tb_book_hero) as book_hero;
+----+-------------+--------------+------------+-------+---------------+-----------------------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+-------+---------------+-----------------------+---------+------+------+----------+-------------+
| 1 | PRIMARY | | NULL | ALL | NULL | NULL | NULL | NULL | 6 | 100.00 | NULL |
| 4 | DERIVED | tb_book_hero | NULL | ALL | NULL | NULL | NULL | NULL | 6 | 100.00 | NULL |
| 3 | SUBQUERY | tb_hero | NULL | index | NULL | idx_book_id_hero_name | 136 | NULL | 6 | 100.00 | Using index |
| 2 | SUBQUERY | tb_book | NULL | index | NULL | uk_book_name | 259 | NULL | 5 | 100.00 | Using index |
+----+-------------+--------------+------------+-------+---------------+-----------------------+---------+------+------+----------+-------------+
4 rows in set, 1 warning (0.00 sec)
mysql> set session optimizer_switch='derived_merge=on';
Query OK, 0 rows affected (0.00 sec)
```
可見，查詢語句中有4個select，先執行的是`select * from tb_book_hero`，然後執行`select count(1) from tb_hero`，再執行`select count(1) from tb_book`，最後執行`select book_count, hero_count from book_hero`
### select_type列
`select_type`表示的是查詢型別，常見的包括`SIMPLE`、`PRIMARY`、`SUBQUERY`、`DERIVED`、`UNION`
(1) SIMPLE：簡單查詢(不包含子查詢和UNION查詢)
```sh
mysql> explain select * from tb_book where book_id = 1;
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| 1 | SIMPLE | tb_book | NULL | const | PRIMARY | PRIMARY | 4 | const | 1 | 100.00 | NULL |
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
```
(2) PRIMARY：複雜查詢中最外層的查詢
(3) SUBQUERY：包含在select中的子查詢(不在from子句中)
(4) DERIVED：包含在from子句中的子查詢，MySQL會將結果存放在一個臨時表中，也稱為派生表(`derived tables`)
這3種select_type見下面的例子
```sh
mysql> set session optimizer_switch='derived_merge=off';
Query OK, 0 rows affected (0.00 sec)
mysql> explain select (select count(1) from tb_book) as book_count from (select * from tb_book_hero) as book_hero;
+----+-------------+--------------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
| 1 | PRIMARY | | NULL | ALL | NULL | NULL | NULL | NULL | 6 | 100.00 | NULL |
| 3 | DERIVED | tb_book_hero | NULL | ALL | NULL | NULL | NULL | NULL | 6 | 100.00 | NULL |
| 2 | SUBQUERY | tb_book | NULL | index | NULL | uk_book_name | 259 | NULL | 5 | 100.00 | Using index |
+----+-------------+--------------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
3 rows in set, 1 warning (0.00 sec)
mysql> set session optimizer_switch='derived_merge=on';
Query OK, 0 rows affected (0.00 sec)
```
(5) UNION：在UNION中的第二個和隨後的select
```sh
mysql> select * from tb_book where book_id = 1 union all select * from tb_book where book_name = '笑傲江湖';
+---------+-----------------------+--------+
| book_id | book_name | author |
+---------+-----------------------+--------+
| 1 | 多情劍客無情劍 | 古龍 |
| 2 | 笑傲江湖 | 金庸 |
+---------+-----------------------+--------+
2 rows in set (0.00 sec)
mysql> explain select * from tb_book where book_id = 1 union all select * from tb_book where book_name = '笑傲江湖';
+----+-------------+---------+------------+-------+---------------+--------------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+--------------+---------+-------+------+----------+-------+
| 1 | PRIMARY | tb_book | NULL | const | PRIMARY | PRIMARY | 4 | const | 1 | 100.00 | NULL |
| 2 | UNION | tb_book | NULL | const | uk_book_name | uk_book_name | 259 | const | 1 | 100.00 | NULL |
+----+-------------+---------+------------+-------+---------------+--------------+---------+-------+------+----------+-------+
2 rows in set, 1 warning (0.00 sec)
```
### table列
`table`表示查詢涉及的表或衍生表。
常見table列是``格式，表示當前查詢依賴`id=N`的查詢，需先執行`id=N`的查詢。上面含`select_type`為`DERIVED`的查詢就是這種情況，這裡不再重複舉例。
### type列
相關文件：
`https://dev.mysql.com/doc/refman/5.7/en/explain-output.html#explain-join-types`
type列是判斷查詢是否高效的重要依據，我們可以通過type欄位的值，判斷此次查詢是`全表掃描`還是`索引掃描`等，進而進一步優化查詢。
一般來說表示查詢效能最優到最差依次為：`NULL > system > const > eq_ref > ref > range > index > ALL`
前面的幾種型別都是利用到了索引來查詢資料, 因此可以過濾部分或大部分資料, 查詢效率自然就比較高了。
而後面的`index`型別的查詢雖然不是全表掃描, 但是它掃描了所有的索引, 因此比`ALL`型別稍快。
所以，應當儘可能地保證查詢達到`range`級別，最好達到`ref`。
(0) NULL: 不用訪問表或者索引，直接就能得到結果，如：在索引列中選取最大值，執行時不需要再訪問表
```sh
mysql> explain select max(book_id) from tb_book;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+------------------------------+
| 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 | Select tables optimized away |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+------------------------------+
1 row in set, 1 warning (0.00 sec)
```
(1) system：The table has only one row. This is a special case of the `const` join type.
當查詢的表只有一行的情況下，`system`是`const`型別的特例，
(2) const：It is used when you compare all parts of a `PRIMARY KEY` or `UNIQUE index` to `constant values`.
針對`主鍵`或`唯一索引`的等值查詢掃描, 最多隻返回一行資料。`const`查詢速度非常快, 因為它僅僅讀取一次即可。
關於type列為`system`、`const`的情況，見下面的示例：
```sh
mysql> set session optimizer_switch='derived_merge=off';
Query OK, 0 rows affected (0.00 sec)
mysql> explain select * from (select * from tb_book where book_id = 5) as book;
+----+-------------+------------+------------+--------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+------------+------------+--------+---------------+---------+---------+-------+------+----------+-------+
| 1 | PRIMARY | | NULL | system | NULL | NULL | NULL | NULL | 1 | 100.00 | NULL |
| 2 | DERIVED | tb_book | NULL | const | PRIMARY | PRIMARY | 4 | const | 1 | 100.00 | NULL |
+----+-------------+------------+------------+--------+---------------+---------+---------+-------+------+----------+-------+
2 rows in set, 1 warning (0.00 sec)
mysql> set session optimizer_switch='derived_merge=on';
Query OK, 0 rows affected (0.00 sec)
```
(3) eq_ref：It is used when all parts of an index are used by the join and the index is a `PRIMARY KEY` or `UNIQUE NOT NULL index`.
此型別通常出現在多表的join查詢，表示對於前表的每一個結果，都只能匹配到後表的一行結果，並且查詢的比較操作通常是`=`，查詢效率較高。
```sh
mysql> select tb_hero.*, tb_book_hero.user_comment from tb_book_hero, tb_hero where tb_book_hero.book_id = 2 and tb_book_hero.hero_id = tb_hero.hero_id;
+---------+-----------+--------------+---------+--------------------------------------------------------------------------------------+
| hero_id | hero_name | skill | book_id | user_comment |
+---------+-----------+--------------+---------+--------------------------------------------------------------------------------------+
| 2 | 令狐沖 | 獨孤九劍 | 2 | 令狐少俠留步! |
| 6 | 任我行 | 吸星大法 | 2 | 有人就有恩怨，有恩怨就有江湖，人心即是江湖，你如何退出！ |
+---------+-----------+--------------+---------+--------------------------------------------------------------------------------------+
2 rows in set (0.00 sec)
mysql> explain select tb_hero.*, tb_book_hero.user_comment from tb_book_hero, tb_hero where tb_book_hero.book_id = 2 and tb_book_hero.hero_id = tb_hero.hero_id;
+----+-------------+--------------+------------+--------+---------------+---------+---------+-----------------------------------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+--------+---------------+---------+---------+-----------------------------------+------+----------+-------+
| 1 | SIMPLE | tb_book_hero | NULL | ref | PRIMARY | PRIMARY | 4 | const | 2 | 100.00 | NULL |
| 1 | SIMPLE | tb_hero | NULL | eq_ref | PRIMARY | PRIMARY | 4 | explain_test.tb_book_hero.hero_id | 1 | 100.00 | NULL |
+----+-------------+--------------+------------+--------+---------------+---------+---------+-----------------------------------+------+----------+-------+
2 rows in set, 1 warning (0.00 sec)
mysql> select tb_hero.*, tb_book_hero.user_comment from tb_book_hero join tb_hero on tb_book_hero.book_id = 2 and tb_book_hero.hero_id = tb_hero.hero_id;
+---------+-----------+--------------+---------+--------------------------------------------------------------------------------------+
| hero_id | hero_name | skill | book_id | user_comment |
+---------+-----------+--------------+---------+--------------------------------------------------------------------------------------+
| 2 | 令狐沖 | 獨孤九劍 | 2 | 令狐少俠留步! |
| 6 | 任我行 | 吸星大法 | 2 | 有人就有恩怨，有恩怨就有江湖，人心即是江湖，你如何退出！ |
+---------+-----------+--------------+---------+--------------------------------------------------------------------------------------+
2 rows in set (0.00 sec)
mysql> explain select tb_hero.*, tb_book_hero.user_comment from tb_book_hero join tb_hero on tb_book_hero.book_id = 2 and tb_book_hero.hero_id = tb_hero.hero_id;
+----+-------------+--------------+------------+--------+---------------+---------+---------+-----------------------------------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+--------+---------------+---------+---------+-----------------------------------+------+----------+-------+
| 1 | SIMPLE | tb_book_hero | NULL | ref | PRIMARY | PRIMARY | 4 | const | 2 | 100.00 | NULL |
| 1 | SIMPLE | tb_hero | NULL | eq_ref | PRIMARY | PRIMARY | 4 | explain_test.tb_book_hero.hero_id | 1 | 100.00 | NULL |
+----+-------------+--------------+------------+--------+---------------+---------+---------+-----------------------------------+------+----------+-------+
2 rows in set, 1 warning (0.00 sec)
```
(4) ref: It is used if the join uses only a leftmost prefix of the key or if the key is not a PRIMARY KEY or UNIQUE index (in other words, if the join cannot select a single row based on the key value).
相比`eq_ref`，不使用唯一索引，而是使用普通索引或者唯一性索引的最左字首，可能會找到多個符合條件的行。
- 簡單的`select`查詢，`author`列上建有普通索引（非唯一索引）
```sh
mysql> explain select * from tb_book where author = '古龍';
+----+-------------+---------+------------+------+---------------+------------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+------------+---------+-------+------+----------+-------+
| 1 | SIMPLE | tb_book | NULL | ref | idx_author | idx_author | 131 | const | 2 | 100.00 | NULL |
+----+-------------+---------+------------+------+---------------+------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
```
- 關聯表查詢，`tb_book_hero`表使用了聯合主鍵`PRIMARY KEY (book_id, hero_id)`，這裡使用到了左邊字首`book_id`進行過濾。
```sh
mysql> explain select * from tb_book_hero where book_id = 3;
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------+
| 1 | SIMPLE | tb_book_hero | NULL | ref | PRIMARY | PRIMARY | 4 | const | 1 | 100.00 | NULL |
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
```
(5) range: It can be used when a key column is compared to a constant using any of the =, <>, >, >=, <, <=, IS NULL, <=>, BETWEEN, LIKE, or IN() operators
掃描部分索引(範圍掃描)，對索引的掃描開始於某一點，返回匹配值域的行，常見於between、<、>、in等查詢
```sh
mysql> explain select * from tb_book where book_id > 3;
+----+-------------+---------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
| 1 | SIMPLE | tb_book | NULL | range | PRIMARY | PRIMARY | 4 | NULL | 2 | 100.00 | Using where |
+----+-------------+---------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
```
(6) index：the index tree is scanned, MySQL can use this type when the query uses only columns that are part of a single index.
表示全索引掃描(full index scan), 和ALL型別類似, 只不過ALL型別是全表掃描, 而index型別則僅僅掃描所有的索引, 而不掃描資料.
```sh
mysql> explain select book_name from tb_book;
+----+-------------+---------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
| 1 | SIMPLE | tb_book | NULL | index | NULL | uk_book_name | 259 | NULL | 5 | 100.00 | Using index |
+----+-------------+---------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
```
上面的例子中, 我們查詢的`book_name`欄位上恰好有索引, 因此我們直接從索引中獲取資料就可以滿足查詢的需求了, 而不需要查詢表中的資料。因此這樣的情況下, type的值是index, 並且Extra的值大多是`Using index`。
(7) ALL: A full table scan is done
表示全表掃描, 這個型別的查詢是效能最差的查詢之一。通常來說, 我們的查詢不應該出現ALL型別的查詢, 因為這樣的查詢在資料量大的情況下, 嚴重降低資料庫的效能。如果一個查詢是ALL型別查詢, 那麼大多可以對相應的欄位新增索引來避免。
```sh
mysql> explain select * from tb_hero where hero_name = '令狐沖';
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| 1 | SIMPLE | tb_hero | NULL | ALL | NULL | NULL | NULL | NULL | 6 | 16.67 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
```
### possible_keys列
表示MySQL在查詢時, 能夠使用到的索引。注意, 即使有些索引在possible_keys中出現, 但是並不表示此索引會真正地被MySQL使用到。MySQL在查詢時具體使用了哪些索引, 由key欄位決定。
### key列
這一列顯示mysql實際採用哪個索引來優化對該表的訪問。如果沒有使用索引，則該列是NULL。
### key_len列
表示查詢優化器使用了索引的位元組數，這個欄位可以評估聯合索引是否完全被使用, 或只有最左部分欄位被使用到。
舉例來說，`tb_hero`表的聯合索引`idx_book_id_hero_name`由`book_id`和`hero_name`兩個列組成，int型別佔4位元組，另外如果欄位允許為NULL，需要1位元組記錄是否為NULL，通過結果中的key_len=5(`tb_hero`.`book_id`允許為NULL)可推斷出查詢使用了第一個列`book_id`列來執行索引查詢；再拿`tb_book_hero`表聯合主鍵`PRIMARY KEY (book_id, hero_id)`舉例，通過key_len=4(`tb_book_hero`.`book_id`不允許為NULL)可推斷出查詢使用了第一個列`book_id`列來執行索引查詢
```sh
mysql> explain select * from tb_hero where book_id = 2;
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
| 1 | SIMPLE | tb_hero | NULL | ref | idx_book_id_hero_name | idx_book_id_hero_name | 5 | const | 2 | 100.00 | NULL |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
mysql> explain select * from tb_book_hero where book_id = 2;
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------+
| 1 | SIMPLE | tb_book_hero | NULL | ref | PRIMARY | PRIMARY | 4 | const | 2 | 100.00 | NULL |
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.01 sec)
```
`key_len`的計算規則如下:
- 字串:
- char(n): n位元組長度
- varchar(n): 如果是`utf8`編碼, 則是`3n + 2`位元組; 如果是`utf8mb4`編碼, 則是`4n + 2`位元組.
- 數值型別:
- TINYINT: 1位元組
- SMALLINT: 2位元組
- MEDIUMINT: 3位元組
- INT: 4位元組
- BIGINT: 8位元組
- 時間型別
- DATE: 3位元組
- TIMESTAMP: 4位元組
- DATETIME: 8位元組
- 欄位屬性:
- NULL屬性佔用一個位元組
- 如果一個欄位是NOT NULL的, 則沒有此屬性
再看下面的計算:
`4 [book_id是int型別] + 1 [book_id允許為NULL] + (4 * 32 + 2) [hero_name是varchar32,且用的是utf8mb4編碼] + 1 [hero_name允許為NULL] = 136`
```sh
mysql> explain select * from tb_hero where book_id = 2 and hero_name = '令狐沖';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-------+
| 1 | SIMPLE | tb_hero | NULL | ref | idx_book_id_hero_name | idx_book_id_hero_name | 136 | const,const | 1 | 100.00 | NULL |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
```
### ref列
> The ref column shows which columns or constants are compared to the index named in the key column to select rows from the table.
顯示的是哪個欄位或常數與key一起被使用
### rows列
MySQL查詢優化器根據統計資訊, 估算SQL要查詢到結果集需要掃描讀取的資料行數，注意這個不是結果集裡的行數。這個值非常直觀顯示SQL的效率好壞, 原則上rows越少越好。
### Extra列
這一列展示的是額外資訊。常見的重要值如下:
(1) Using index
表示查詢在索引樹中就可查詢所需資料, 不用掃描表資料檔案, 往往說明效能不錯
```sh
mysql> explain select hero_id from tb_book_hero where book_id = 2;
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------------+
| 1 | SIMPLE | tb_book_hero | NULL | ref | PRIMARY | PRIMARY | 4 | const | 2 | 100.00 | Using index |
+----+-------------+--------------+------------+------+---------------+---------+---------+-------+------+----------+-------------+
1 row in set, 1 warning (0.01 sec)
mysql> explain select book_id from tb_book where author = '金庸';
+----+-------------+---------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
| 1 | SIMPLE | tb_book | NULL | ref | idx_author | idx_author | 131 | const | 3 | 100.00 | Using index |
+----+-------------+---------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
```
(2) Using where
查詢的列沒有全部被索引覆蓋
```sh
mysql> explain select book_id, book_name from tb_book where author = '金庸';
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| 1 | SIMPLE | tb_book | NULL | ALL | idx_author | NULL | NULL | NULL | 5 | 60.00 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
```
(3) Using temporary
查詢有使用臨時表，一般出現於排序、分組、多表join、distinct查詢等等。
舉例子如下：`tb_book`表對`book_name`欄位建立了唯一性索引，這時候distinct查詢Extra列為`Using index`; `tb_hero`表的`skill`欄位上沒有任何索引，這時候distinct查詢Extra列為`Using temporary`
```sh
mysql> select distinct book_name from tb_book;
+-----------------------+
| book_name |
+-----------------------+
| 倚天屠龍記 |
| 多情劍客無情劍 |
| 射鵰英雄傳 |
| 笑傲江湖 |
| 絕代雙驕 |
+-----------------------+
5 rows in set (0.00 sec)
mysql> explain select distinct book_name from tb_book;
+----+-------------+---------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
| 1 | SIMPLE | tb_book | NULL | index | uk_book_name | uk_book_name | 259 | NULL | 5 | 100.00 | Using index |
+----+-------------+---------+------------+-------+---------------+--------------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
mysql> select distinct skill from tb_hero;
+-----------------+
| skill |
+-----------------+
| 小李飛刀 |
| 獨孤九劍 |
| 九陽神功 |
| 降龍十八掌 |
| 移花接玉 |
| 吸星大法 |
+-----------------+
6 rows in set (0.00 sec)
mysql> explain select distinct skill from tb_hero;
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-----------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-----------------+
| 1 | SIMPLE | tb_hero | NULL | ALL | NULL | NULL | NULL | NULL | 6 | 100.00 | Using temporary |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-----------------+
1 row in set, 1 warning (0.00 sec)
```
(4) Using filesort
表示MySQL不能通過索引順序達到排序效果，需額外的排序操作，資料較小時在記憶體排序，否則需要在磁碟完成排序。這種情況下一般也是要考慮使用索引來優化的。
```sh
mysql> explain select book_id, hero_name from tb_hero order by hero_name;
+----+-------------+---------+------------+-------+---------------+-----------------------+---------+------+------+----------+-----------------------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+-----------------------+---------+------+------+----------+-----------------------------+
| 1 | SIMPLE | tb_hero | NULL | index | NULL | idx_book_id_hero_name | 136 | NULL | 6 | 100.00 | Using index; Using filesort |
+----+-------------+---------+------------+-------+---------------+-----------------------+---------+------+------+----------+-----------------------------+
1 row in set, 1 warning (0.00 sec)
mysql> explain select book_id, hero_name from tb_hero order by book_id, hero_name;
+----+-------------+---------+------------+-------+---------------+-----------------------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+-----------------------+---------+------+------+----------+-------------+
| 1 | SIMPLE | tb_hero | NULL | index | NULL | idx_book_id_hero_name | 136 | NULL | 6 | 100.00 | Using index |
+----+-------------+---------+------------+-------+---------------+-----------------------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
```
`tb_hero`表上有聯合索引`INDEX idx_book_id_hero_name(book_id, hero_name) USING BTREE`
但是`order by hero_name`, 不能使用索引進行優化(下一篇部落格會介紹聯合索引的結構), 進而會產生`Using filesort`
如果將排序依據改為`order by book_id, hero_name`, 就不會出現`Using filesort`了。
(5) Select tables optimized away
比如下面的例子：
```sh
mysql> explain select min(book_id), max(book_id) from tb_book;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+------------------------------+
| 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 | Select tables optimized away |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+------------------------------+
1 row in set, 1 warning (0.00