我有两张桌子A和B.
表A
guid | username
---------------
A | name1
B | name2
C | name3
使用索引
+--------+------------+-------------------------------+--------------+--------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table | Non_unique | Key_name | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+--------+------------+-------------------------------+--------------+--------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| tableA | 0 | PRIMARY | 1 | guid | A | 2242900 | NULL | NULL | | BTREE | | |
+--------+------------+-------------------------------+--------------+--------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
表B
guidA | guidB
-------------
A | C
B | C
使用索引
+----------+------------+---------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table | Non_unique | Key_name | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+----------+------------+---------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| tableB | 1 | guidA_idx | 1 | userid | A | 3 | NULL | NULL | | BTREE | | |
| tableB | 1 | guidB_idx | 1 | userid | A | 3 | NULL | NULL | | BTREE | | |
+----------+------------+---------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
我想要的是在给定 guidB 的情况下进行查询我获取相应 {{1}的 username }
仅从表A中选择其快速..
guidA使用DESC
select guid from tableA where guid IN ('A', 'B');
从表B中选择也很快......（它只有三行，所以没有使用索引）
+----+-------------+---------+-------+---------------+---------+---------+------+------+--------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------+-------+---------------+---------+---------+------+------+--------------------------+
| 1 | SIMPLE | tableA | range | PRIMARY | PRIMARY | 66 | NULL | 3 | Using where; Using index |
+----+-------------+---------+-------+---------------+---------+---------+------+------+--------------------------+
DESC
select guidA from tableB where guidB = 'C';
然而！如果我做
+----+-------------+--------+------+--------------+------+---------+------+------+-------------+
| id | select_type | table | type | possible_keys| key | key_len | ref | rows | Extra |
+----+-------------+--------+------+--------------+------+---------+------+------+-------------+
| 1 | SIMPLE | tableB | ALL | guidB_idx | NULL | NULL | NULL | 3 | Using where |
+----+-------------+--------+------+--------------+------+---------+------+------+-------------+
我得到了正确的结果，但查询速度很慢，因为tableA中的guid索引从未使用过。
DESC
select guid, username from tableA where guid IN (select guidA from tableB where guidB = 'C');
正如您所看到的，它可以进行全表扫描。我在某处读到使用+----+-------------+--------+------+---------------------+------+---------+------+---------+-------------------------------------------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+--------+------+---------------------+------+---------+------+---------+-------------------------------------------------------------------------+
| 1 | SIMPLE | tableA | ALL | NULL | NULL | NULL | NULL | 2242900 | NULL |
| 1 | SIMPLE | tableB | ALL | guidA_idx,guidB_idx | NULL | NULL | NULL | 3 | Using where; FirstMatch(tableA); Using join buffer (Block Nested Loop) |
+----+-------------+--------+------+---------------------+------+---------+------+---------+-------------------------------------------------------------------------+
有优化问题，但重写查询为 WHERE IN (subquery) 并没有给出不同的结果。我总是在LEFT JOIN, RIGHT JOIN, or INNER JOIN上进行全表扫描。我每次都得到正确的结果。
加入...
tableA我的MySQL版本 Ver 14.14 Distrib 5.6.28，for Win64（x86_64）
答案 0 :(得分：1)
您的列必须具有相同的类型和排序规则。
不同的排序规则或不同类型将避免使用任何索引。