文章目录
- 1、前置准备
- 2、ChatGPT的答案
- 3、实践证明
- SQL1
- SQL2
- SQL3
- SQL4
- SQL5
- 4、总结
1、前置准备
首先创建我们要测试的库表
CREATE TABLE `lianhe_index` (`id` int(11) NOT NULL AUTO_INCREMENT COMMENT 'id',`name` varchar(255) DEFAULT NULL,`age` int(11) DEFAULT NULL,`number` int(11) DEFAULT NULL,`create_time` datetime DEFAULT NULL ON UPDATE CURRENT_TIMESTAMP,PRIMARY KEY (`id`),KEY `idx_nameagenumber` (`name`,`age`,`number`) USING BTREE
) ENGINE=InnoDB AUTO_INCREMENT=200001 DEFAULT CHARSET=utf8;
主要建立了主键索引id和联合索引idx_nameagenumber，造了10W条数据。
一般函数可以使用在select子句中的列，也可以使用在where子句中的列，下面我们就这两种情况来分析下。
2、ChatGPT的答案
首先我们来问下ChatGPT
在MySQL中，使用函数作为select字段通常不会走索引。这是因为函数的使用会导致MySQL无法直接匹配索引键值，而是需要对所有记录进行函数计算后才能确定结果。然而，某些特殊情况下，MySQL可能会通过某些优化策略来利用索引，例如在某些情况下可以使用覆盖索引来避免访问表数据。
对于where字段的情况，使用函数作为条件通常也不会走索引。同样的原因，函数的使用会导致MySQL无法直接匹配索引键值，而是需要对所有记录进行函数计算后才能确定结果。因此，如果希望where字段能够走索引，建议尽量避免使用函数。
上面是ChatGPT的回答，回答正确？先保留答案，接下来我们来实践下~
3、实践证明
我们主要分析下面这5条SQL，先猜猜看会不会走索引？
SELECT id,age,number,UPPER(name),create_time FROM lianhe_index where name='林伟' and age=69 and number=76291;
SELECT id,age,number,name,create_time FROM lianhe_index where UPPER(name)='伟A' and age=69 and number=76291;
SELECT id,age,number,name,create_time FROM lianhe_index where name=UPPER(name) and age=69 and number=76291;
SELECT id,age,number,name,create_time FROM lianhe_index where UPPER(name)='伟A' and age=69 and number=76291 and DATE(create_time)="2022-11-11";
SELECT id,age,number,name,create_time FROM lianhe_index where name=UPPER(name) and age=69 and number=76291 and DATE(create_time)="2022-11-11";
执行计划如下，和你想的一样？
mysql> EXPLAIN SELECT id,age,number,UPPER(name),create_time FROM lianhe_index where name='林伟' and age=69 and number=76291;
EXPLAIN SELECT id,age,number,name,create_time FROM lianhe_index where UPPER(name)='伟A' and age=69 and number=76291;
EXPLAIN SELECT id,age,number,name,create_time FROM lianhe_index where name=UPPER(name) and age=69 and number=76291;
EXPLAIN SELECT id,age,number,name,create_time FROM lianhe_index where UPPER(name)='伟A' and age=69 and number=76291 and DATE(create_time)="2022-11-11";
EXPLAIN SELECT id,age,number,name,create_time FROM lianhe_index where name=UPPER(name) and age=69 and number=76291 and DATE(create_time)="2022-11-11";
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------+
| 1 | SIMPLE | lianhe_index | NULL | ref | idx_nameagenumber | idx_nameagenumber | 778 | const,const,const | 1 | 100.00 | NULL |
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------+
1 row in set (0.02 sec)+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 1.00 | Using where |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
1 row in set (0.02 sec)+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 0.10 | Using where |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
1 row in set (0.02 sec)+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 1.00 | Using where |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
1 row in set (0.02 sec)+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 0.10 | Using where |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
1 row in set (0.02 sec)下面我们逐个来分析下
SQL1
SELECT id,age,number,UPPER(name),create_time FROM lianhe_index where name='林伟' and age=69 and number=76291;
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------+
| 1 | SIMPLE | lianhe_index | NULL | ref | idx_nameagenumber | idx_nameagenumber | 778 | const,const,const | 1 | 100.00 | NULL |
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------+
可以看到，执行计划中的key是使用到了组合索引idx_nameagenumber，扫码的行数rows=1，所以，索引生效
虽然RIGHT(name,1)是一个函数，但是它是在where子句中应用于已经在索引中的列name上的，所以MySQL可以使用索引来定位匹配的行
如果我们不查询create_time，还会发现执行计划中Extra中出现了Using index，表示使用了覆盖索引
SELECT id,age,number,UPPER(name) FROM lianhe_index where name='林伟' and age=69 and number=76291;
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ref | idx_nameagenumber | idx_nameagenumber | 778 | const,const,const | 1 | 100.00 | Using index |
+----+-------------+--------------+------------+------+-------------------+-------------------+---------+-------------------+------+----------+-------------+
SQL2
SELECT id,age,number,name,create_time FROM lianhe_index where UPPER(name)='伟A' and age=69 and number=76291;
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 1.00 | Using where |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
可以看到，执行计划中的key=null，rows=99954，Extra列显示为"Using where"，这表示查询使用了where子句进行过滤，但没有使用索引
如果我们不查询create_time，可以发现在执行计划中，“possible_keys"列显示为"NULL”，这表示没有其他可用的索引。但是在"key"列中显示了"idx_nameagenumber"，这表示该索引被选择用于执行查询。此外，Extra列显示了"Using where; Using index"，这表示查询使用了where子句进行过滤，并且使用了索引来定位匹配的行。所以是使用了索引，但效果不佳，个人觉得执行计划这里是有问题的
SELECT id,age,number,name FROM lianhe_index where UPPER(name)='伟A' and age=69 and number=76291;
+----+-------------+--------------+------------+-------+---------------+-------------------+---------+------+-------+----------+--------------------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+-------+---------------+-------------------+---------+------+-------+----------+--------------------------+
| 1 | SIMPLE | lianhe_index | NULL | index | NULL | idx_nameagenumber | 778 | NULL | 99954 | 1.00 | Using where; Using index |
+----+-------------+--------------+------------+-------+---------------+-------------------+---------+------+-------+----------+--------------------------+
SQL3
SELECT id,age,number,name,create_time FROM lianhe_index where name=UPPER(name) and age=69 and number=76291;
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 0.10 | Using where |
可以看到，执行计划中的key=null，rows=99954，Extra列显示为"Using where"，这表示查询使用了where子句进行过滤，但没有使用索引
SQL4
SELECT id,age,number,name,create_time FROM lianhe_index where UPPER(name)='伟A' and age=69 and number=76291 and DATE(create_time)="2022-11-11";
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 1.00 | Using where |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
可以看到，执行计划中的key=null，rows=99954，Extra列显示为"Using where"，这表示查询使用了where子句进行过滤，但没有使用索引
SQL5
SELECT id,age,number,name,create_time FROM lianhe_index where name=UPPER(name) and age=69 and number=76291 and DATE(create_time)="2022-11-11";
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+--------------+------------+------+---------------+------+---------+------+-------+----------+-------------+
| 1 | SIMPLE | lianhe_index | NULL | ALL | NULL | NULL | NULL | NULL | 99954 | 0.10 | Using where |
可以看到，执行计划中的key=null，rows=99954，Extra列显示为"Using where"，这表示查询使用了where子句进行过滤，但没有使用索引
4、总结
当函数应用于select子句中的索引列时，索引仍然生效；当函数应用于where子句中的索引列时，MySQL可能无法使用索引来加快查询速度，虽然存在部分情况，执行计划显示keys使用了索引，但实际效果不佳，可以当作索引失效了。
![[C++随想录] 继承](https://img-blog.csdnimg.cn/40a7d59d8b9d4e298b509dbe25292f83.png)
