前言
近期多次聊到sql_mode的话题，也是多次遇到相关问题，今天就趁热打铁，再给大家带来一个sql_mode的案例分享。
场景模拟
基于业务敏感性的考虑，下面涉及的表、存储过程等均非真实数据，但并不影响排查过程。
（1）客户侧开发童鞋创建了一个存储过程，该存储过程没有严格遵守group by标准语法
session 1：
mysql> delimiter //
mysql> create procedure test_for_group_by()
-> begin
-> select k,pad,count(*) from test.test group by k;
-> end //
Query OK, 0 rows affected (0.01 sec)
mysql> delimiter ;
（2）客户侧开发童鞋调用该存储过程，报错ERROR 1140；因为当时存储过程比较复杂，改造起来比较麻烦，所以客户侧选择修改sql_mode
session 1： mysql> call test_for_group_by(); ERROR 1140 (42000): In aggregated query without GROUP BY, expression #1 of SELECT list contains nonaggregated column 'test.test.k'; this is incompatible with sql_mode=only_full_group_by
（3）客户侧修改完sql_mode，再次执行，发现仍然报错ERROR 1140
session 2： mysql> set global sql_mode='STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION'; Query OK, 0 rows affected (0.00 sec) session 1： mysql> call test_for_group_by(); ERROR 1140 (42000): In aggregated query without GROUP BY, expression #1 of SELECT list contains nonaggregated column 'test.test.k'; this is incompatible with sql_mode=only_full_group_by
（4）此时想到，修改系统变量，只对新建连接有效，对已有连接不起作用；于是，让客户侧重新建立连接，确认系统变量已生效，再次调用存储过程，但仍然报错ERROR 1140，重复尝试几次都是这个结果
session 3： mysql> show variables like 'sql_mode'; +---------------+------------------------------------------------------------------------------------------------------------------------+ | Variable_name | Value | +---------------+------------------------------------------------------------------------------------------------------------------------+ | sql_mode | STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION | +---------------+------------------------------------------------------------------------------------------------------------------------+ 1 row in set (0.01 sec) mysql> call test_for_group_by(); ERROR 1140 (42000): In aggregated query without GROUP BY, expression #1 of SELECT list contains nonaggregated column 'test.test.k'; this is incompatible with sql_mode=only_full_group_by
（5）进一步排查，让客户侧在该会话，执行非标准的group by语句，发现可以正常执行
session 3： mysql> select user,host,count(*) From mysql.user group by user; +---------------+-----------+----------+ | user | host | count(*) | +---------------+-----------+----------+ | mysql.session | localhost | 1 | | mysql.sys | localhost | 1 | | root | localhost | 1 | | rpl_user | % | 1 | | test | % | 1 | +---------------+-----------+----------+ 5 rows in set (0.00 sec)
（6）继续排查发现，该存储过程的sql_mode，还是包括ONLY_FULL_GROUP_BY，因此执行报错
session 2： mysql> select routine_catalog,routine_schema,routine_name,routine_type,created,last_altered,sql_mode from routines where routine_name='test_for_group_by'; +-----------------+----------------+-------------------+--------------+---------------------+---------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ | routine_catalog | routine_schema | routine_name | routine_type | created | last_altered | sql_mode | +-----------------+----------------+-------------------+--------------+---------------------+---------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ | def | test | test_for_group_by | PROCEDURE | 2020-12-24 12:12:10 | 2020-12-24 12:12:10 | ONLY_FULL_GROUP_BY,STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION | +-----------------+----------------+-------------------+--------------+---------------------+---------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ 1 row in set (0.00 sec)
（7）这里我们也可以知道，系统变量修改只对新建对象有效，对已有对象不生效；解决办法很简单，重建该存储过程即可
session 3：
mysql> drop procedure test_for_group_by;
Query OK, 0 rows affected (0.01 sec)
mysql> delimiter //
mysql> create procedure test_for_group_by()
-> begin
-> select k,pad,count(*) from test.test group by k;
-> end //
Query OK, 0 rows affected (0.01 sec)
mysql> delimiter ;
mysql> call test_for_group_by();
+--------+-------------------------------------------------------------+----------+
| k | pad | count(*) |
+--------+-------------------------------------------------------------+----------+
| 393975 | 35227182905-15234265621-59793845249-15413569710-23749555118 | 1 |
| 495688 | 09512147864-77936258834-40901700703-13541171421-15205431759 | 1 |
| 497896 | 13152283289-69561545685-52868757241-04245213425-69280254356 | 1 |
| 498573 | 43131080328-59298106536-35954612339-97546855884-75769514803 | 1 |
| 500775 | 27590239742-20204899609-34345212327-79811525340-24267764271 | 1 |
| 501885 | 63188288836-92351140030-06390587585-66802097351-49282961843 | 1 |
| 503330 | 01495266405-82925129145-92643983850-90243995398-18709399387 | 1 |
| 503666 | 40929980986-33813039690-13155419391-97985458477-39771362212 | 1 |
| 504353 | 00505722282-72931248925-57037623248-81117963809-88658076981 | 1 |
| 514246 | 21979564480-87492594656-60524686334-78820761788-57684966682 | 1 |
+--------+-------------------------------------------------------------+----------+
10 rows in set (0.00 sec)
Query OK, 0 rows affected (0.00 sec)
总结
通过这个案例，我们可以知道，修改sql_mode系统变量，只对新建连接和新建对象(主要包括函数和存储过程)有效，对已有连接和已有对象不生效。
以上就是MySQL sql_mode修改不生效的原因及解决的详细内容，更多关于MySQL sql_mode修改不生效的资料请关注服务器之家其它相关文章！