从这个SQL语句中,我试图理解pkey的用法,我们已经有了复合键作为主键,请有人解释一下。
CREATE TABLE categories_products (
category_id int unsigned not null,
product_id int unsigned not null,
PRIMARY KEY (category_id, product_id),
KEY pkey (product_id),
FOREIGN KEY (category_id) REFERENCES categories (id)
ON DELETE CASCADE
ON UPDATE CASCADE,
FOREIGN KEY (product_id) REFERENCES products (id)
ON DELETE CASCADE
ON UPDATE CASCADE
);发布于 2015-05-21 05:27:03
通过扩展解释,并从您的例子出发:
首先,让我们设置环境:
mysql>
mysql> create table categories_products (
-> category_id int unsigned not null,
-> product_id int unsigned not null,
-> somefield varchar(25),
-> primary key (category_id, product_id)
-> );
Query OK, 0 rows affected (0.10 sec)
mysql>
mysql> insert into categories_products(category_id, product_id) values (1, 1), (1, 2), (2, 2);
Query OK, 3 rows affected (0.05 sec)
Records: 3 Duplicates: 0 Warnings: 0我删除了外键,因为它们对此没有任何影响,我还添加了一个额外的字段,因为它使结果更容易解释。我会在那之后再谈。
首先,我们尝试查询一个特定类别:
mysql> explain select * from categories_products where category_id = 1;
+----+-------------+---------------------+------+---------------+---------+---------+-------+------+-------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------------------+------+---------------+---------+---------+-------+------+-------+
| 1 | SIMPLE | categories_products | ref | PRIMARY | PRIMARY | 4 | const | 2 | |
+----+-------------+---------------------+------+---------------+---------+---------+-------+------+-------+
1 row in set (0.00 sec)从结果可以看出,它确实使用了PRIMARY键,并且只检查2行,因为索引知道这是匹配where约束的全部内容。
我们的下一个测试同时查询产品和类别:
mysql> explain select * from categories_products where category_id = 1 and product_id = 1;
+----+-------------+---------------------+-------+---------------+---------+---------+-------------+------+-------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------------------+-------+---------------+---------+---------+-------------+------+-------+
| 1 | SIMPLE | categories_products | const | PRIMARY | PRIMARY | 8 | const,const | 1 | |
+----+-------------+---------------------+-------+---------------+---------+---------+-------------+------+-------+
1 row in set (0.00 sec)我们再一次看到主键正在被使用,但这一次更好的是,它只返回一行。
接下来,让我们来看看一个产品:
mysql> explain select * from categories_products where product_id = 1;
+----+-------------+---------------------+------+---------------+------+---------+------+------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------------------+------+---------------+------+---------+------+------+-------------+
| 1 | SIMPLE | categories_products | ALL | NULL | NULL | NULL | NULL | 3 | Using where |
+----+-------------+---------------------+------+---------------+------+---------+------+------+-------------+
1 row in set (0.00 sec)这一次,查询无法找到合适的索引来使用,因此不得不使用where缩小结果。这比它能够执行索引查找的效率要低。
为什么是这种情况?为什么优化者可以使用category_id,而不能使用product_id,即使它们都在综合指数中?因为MySQL从左到右读取索引。考虑一个大型的综合指数,(f1, f2, f3, f4)。这意味着您可以访问这些附加索引,即(f1)、(f1, f2)、(f1, f2, f3)。
现在,让我们将您的pkey索引添加到其中,看看会发生什么。
mysql> create index pkey on categories_products(product_id);
explain select * from categories_products where product_id = 1;
Query OK, 0 rows affected (0.17 sec)
Records: 0 Duplicates: 0 Warnings: 0
mysql>
mysql> explain select * from categories_products where product_id = 1;
+----+-------------+---------------------+------+---------------+------+---------+-------+------+-------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------------------+------+---------------+------+---------+-------+------+-------+
| 1 | SIMPLE | categories_products | ref | pkey | pkey | 4 | const | 1 | |
+----+-------------+---------------------+------+---------------+------+---------+-------+------+-------+现在,我们可以使用索引来执行与以前相同的查询。
现在来说说我为什么要增加一个字段。当主键覆盖整个表时,就像在本例中所做的那样,任何查询都将从索引中读取整个结果,而不是从表中读取,这可能会在解释结果中产生一些误导条目。例如,如果删除该附加字段和pkey索引并重新运行查询,结果如下:
mysql> explain select * from categories_products where product_id = 1;
+----+-------------+---------------------+-------+---------------+---------+---------+------+------+--------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------------------+-------+---------------+---------+---------+------+------+--------------------------+
| 1 | SIMPLE | categories_products | index | NULL | PRIMARY | 8 | NULL | 3 | Using where; Using index |
+----+-------------+---------------------+-------+---------------+---------+---------+------+------+--------------------------+
1 row in set (0.00 sec)现在乍一看,主索引似乎正在被使用,但仔细观察,我们发现它仍然在检查整个表(3行),并限制了where和索引的结果。
无论如何,这就是为什么您需要在product_id上建立索引。
https://stackoverflow.com/questions/30364846
复制相似问题