mysql实战之 批量update

mysql实战之批量update

现阶段我们的业务量很小，要对admin_user表中的relationship字段进行更新，指定id是409。已知409是公司内的一服务中心，需要把该服务中心放到区代下面， 区代id是3486，相关字段是paren_id（父子级别关系），需要把409已经409旗下所有商家以及客户都转移到3486下面，这个对应字段是relationship ，业务介绍完毕，下面来实战操作：

查看各个id现在的relationship情况

mysql> SELECT  relationship FROM admin_user WHERE parent_id=409;

+-------------------------+

| relationship            |

+-------------------------+

| ,1,30,304,405,409,437,  |

| ,1,30,304,405,409,450,  |

| ,1,30,304,405,409,464,  |

| ,1,30,304,405,409,465,  |

| ,1,30,304,405,409,471,  |

| ,1,30,304,405,409,505,  |

| ,1,30,304,405,409,540,  |

| ,1,30,304,405,409,793,  |

| ,1,30,304,405,409,794,  |

| ,1,30,304,405,409,1801, |

| ,1,30,304,405,409,1802, |

| ,1,30,304,405,409,1949, |

| ,1,30,304,405,409,2015, |

+-------------------------+

13 rows in set (0.02 sec)

mysql> SELECT relationship FROM admin_user WHERE parent_id=3486;

+---------------------+

| relationship        |

+---------------------+

| ,1,30,304,3486,409, |

+---------------------+

1 row in set (0.02 sec)

现在就是要把parent_id=409 的relationship中’,1,30,304,405,409,437,’所有的409换成3486，

笨方法：一条一条的update，但是业务量小，这样可以实现，假如是几万条数据的话这样操作是不可取的，还有手动更改某一个数很难保证正确性。

UPDATE admin_user SET relationship=',1,30,304,405,3486,437,' WHERE id=437;

UPDATE admin_user SET relationship=',1,30,304,405,3486,450,' WHERE id=450;

UPDATE admin_user SET relationship=',1,30,304,405,3486,464,' WHERE id=464;

。。。。。。。。。。。

其实我就是用这样的方法去更新的，后来经过领导指点采取新的方法实现一条语句更新来实现业务，

使用到的是连接函数concat，查看语法

mysql> help concat

Name: 'CONCAT'

Description:

Syntax:

CONCAT(str1,str2,...)

Returns the string that results from concatenating the arguments. May

have one or more arguments. If all arguments are nonbinary strings, the

result is a nonbinary string. If the arguments include any binary

strings, the result is a binary string. A numeric argument is converted

to its equivalent nonbinary string form.

CONCAT() returns NULL if any argument is NULL.

URL: http://dev.mysql.com/doc/refman/5.6/en/string-functions.html

Examples:

mysql> SELECT CONCAT('My', 'S', 'QL');

-> 'MySQL'

mysql> SELECT CONCAT('My', NULL, 'QL');

-> NULL

mysql> SELECT CONCAT(14.3);

-> '14.3'

我需要替换掉最左边之前的字符，都是固定的，先写下查询的sql语句试试，

mysql> SELECT CONCAT(',1,30,304,3486,409,',RIGHT(relationship,LENGTH(relationship)-LENGTH(',1,30,304,405,409,'))) AS r1,relationship FROM admin_user WHERE relationship LIKE ',1,30,304,405,409,%';

+--------------------------+-------------------------+

| r1                       | relationship            |

+--------------------------+-------------------------+

| ,1,30,304,3486,409,437,  | ,1,30,304,405,409,437,  |

| ,1,30,304,3486,409,450,  | ,1,30,304,405,409,450,  |

| ,1,30,304,3486,409,464,  | ,1,30,304,405,409,464,  |

| ,1,30,304,3486,409,465,  | ,1,30,304,405,409,465,  |

| ,1,30,304,3486,409,471,  | ,1,30,304,405,409,471,  |

| ,1,30,304,3486,409,505,  | ,1,30,304,405,409,505,  |

| ,1,30,304,3486,409,540,  | ,1,30,304,405,409,540,  |

| ,1,30,304,3486,409,793,  | ,1,30,304,405,409,793,  |

| ,1,30,304,3486,409,794,  | ,1,30,304,405,409,794,  |

| ,1,30,304,3486,409,1801, | ,1,30,304,405,409,1801, |

| ,1,30,304,3486,409,1802, | ,1,30,304,405,409,1802, |

| ,1,30,304,3486,409,1949, | ,1,30,304,405,409,1949, |

| ,1,30,304,3486,409,2015, | ,1,30,304,405,409,2015, |

+--------------------------+-------------------------+

13 rows in set (0.01 sec)

很显然，r1就是我想要的结果，下面可以更改成一个更新语句了，

mysql> update admin_user set relationship=concat(',1,30,304,3486,409,',RIGHT(relationship,LENGTH(relationship)-LENGTH(',1,30,304,405,409,'))) where relationship LIKE ',1,30,304,405,409,%';

Query OK, 13 rows affected (0.05 sec)

Rows matched: 13  Changed: 13  Warnings: 0

mysql> select relationship from admin_user where parent_id=409;

+--------------------------+

| relationship             |

+--------------------------+

| ,1,30,304,3486,409,437,  |

| ,1,30,304,3486,409,450,  |

| ,1,30,304,3486,409,464,  |

| ,1,30,304,3486,409,465,  |

| ,1,30,304,3486,409,471,  |

| ,1,30,304,3486,409,505,  |

| ,1,30,304,3486,409,540,  |

| ,1,30,304,3486,409,793,  |

| ,1,30,304,3486,409,794,  |

| ,1,30,304,3486,409,1801, |

| ,1,30,304,3486,409,1802, |

| ,1,30,304,3486,409,1949, |

| ,1,30,304,3486,409,2015, |

+--------------------------+

13 rows in set (0.02 sec)

搞定，期间我尝试过用when case来实现这个业务，跟单个update一样繁琐，就这13条数据也看不出来有多效率，所以还是这个拼接替换好用，菜鸟级水平奉上。