數據庫版本(select version();): 5.7.32-log
sakila 下載地址: https://dev.mysql.com/doc/index-other.html
type 表示 MySQL 在表中找到所需行的方式, 又稱"訪問類型".
ALL, index, range, ref, eq_ref, const, system, NULL (從左到右, 性能從差到好)
1. ALL
全表掃描(Full Table Scan), MySQL將遍歷全表以找到匹配的行.
mysql> explain select * from film where rating ='G';
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
| 1 | SIMPLE | film | NULL | ALL | NULL | NULL | NULL | NULL | 1000 | 20.00 | Using where |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
film 表中 rating 字段沒有索引.
2. index
全索引掃描(Full Index Scan), index 與 ALL 區別為 index 類型只遍歷索引樹. MYSQL 遍歷整個索引來查找匹配的行.
mysql> explain select title from film;
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| 1 | SIMPLE | film | NULL | index | NULL | idx_title | 514 | NULL | 1000 | 100.00 | Using index |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
雖然 where 條件中沒有用到索引, 但是要取出的列 title 是索引包含的列, 所以只要全掃描 title 索引即可, 直接使用索引樹查找數據.
3. range
索引范圍掃描, 常見於 '<', '<=', '>', '>=', 'between' 等操作符.
mysql> explain select * from film where film_id > 100;
+----+-------------+-------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+-------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
| 1 | SIMPLE | film | NULL | range | PRIMARY | PRIMARY | 2 | NULL | 900 | 100.00 | Using where |
+----+-------------+-------+------------+-------+---------------+---------+---------+------+------+----------+-------------+
因為 film_id 是索引, 所以只要查找索引的某個范圍即可, 通過索引找到具體的數據.
4. ref
使用非唯一性索引或者唯一索引的前綴掃描, 返回匹配某個單獨值的記錄行.
mysql> explain select * from payment where customer_id = 10;
+----+-------------+---------+------------+------+--------------------+--------------------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+--------------------+--------------------+---------+-------+------+----------+-------+
| 1 | SIMPLE | payment | NULL | ref | idx_fk_customer_id | idx_fk_customer_id | 2 | const | 25 | 100.00 | NULL |
+----+-------------+---------+------------+------+--------------------+--------------------+---------+-------+------+----------+-------+
customer_id 在 payment 表中是非唯一性索引
5. eq_ref
類似ref, 區別就在使用的索引是唯一索引. 在聯表查詢中使用 primary key 或者 unique key 作為關聯條件.
mysql> explain select * from film a left join film_text b on a.film_id = b.film_id;
+----+-------------+-------+------------+--------+---------------+---------+---------+------------------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+-------+------------+--------+---------------+---------+---------+------------------+------+----------+-------------+
| 1 | SIMPLE | a | NULL | ALL | NULL | NULL | NULL | NULL | 1000 | 100.00 | NULL |
| 1 | SIMPLE | b | NULL | eq_ref | PRIMARY | PRIMARY | 2 | sakila.a.film_id | 1 | 100.00 | Using where |
+----+-------------+-------+------------+--------+---------------+---------+---------+------------------+------+----------+-------------+
6. const/system
當 MySQL 對查詢某部分進行優化, 並轉換為一個常量時, 使用這些類型訪問. 如將主鍵置於 where 列表中, MySQL 就能將該查詢轉換為一個常量, system 是 const 類型的特例, 當查詢的表只有一行的情況下使用 system.
mysql> explain select * from film where film_id = 1;
+----+-------------+-------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+-------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| 1 | SIMPLE | film | NULL | const | PRIMARY | PRIMARY | 2 | const | 1 | 100.00 | NULL |
+----+-------------+-------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
7. NULL
MySQL 不用訪問表或者索引就直接能到結果.
mysql> explain select 1 from dual where 1;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------+
| 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 | No tables used |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------+
dual是一個虛擬的表, 可以直接忽略.