?
?通常情况下,一个我们在做一个产品的时候,一开始可能由于设计考虑不周或者程序写的不够严谨,某个字段上的值产生重复了,但是又必须去掉,这个时候就稍微麻烦了一点,直接加一个?UNIQUE KEY
?肯定是不行了,因为会报错。
?现在,我们来采用一种变通的办法,不过可能会丢失一些数据 :)
?在这里,我们设定一个表,其结构如下:
mysql> desc `user`; +-------+------------------+------+-----+---------+----------------+ | Field | Type | Null | Key | Default | Extra | +-------+------------------+------+-----+---------+----------------+ | id | int(10) unsigned | NO | PRI | NULL | auto_increment | | name | char(10) | NO | | | | | extra | char(10) | NO | | | | +-------+------------------+------+-----+---------+----------------+
?原来表中的数据假定有以下几条:
mysql> SELECT * FROM `user`; +----+-------+--------+ | id | name | extra | +----+-------+--------+ | 1 | user1 | user1 | | 2 | user2 | user2 | | 3 | user3 | user3 | | 4 | user4 | user4 | | 5 | user5 | user5 | | 6 | user3 | user6 | | 7 | user6 | user7 | | 8 | user2 | user8 | | 9 | USER2 | user9 | | 10 | USER6 | user10 | +----+-------+--------+
1、将原来的数据导出
mysql>SELECT * INTO OUTFILE '/tmp/user.txt' FROM `user`;
2、清空数据表
mysql>TRUNCATE TABLE `user`;
3、创建唯一索引,并且修改 `name` 字段的类型为?BINARY CHAR
?区分大小写
mysql> ALTER TABLE `user` MODIFY `name` CHAR(10) BINARY NOT NULL DEFAULT ''; mysql> ALTER TABLE `user` ADD UNIQUE KEY ( `name` );
现在来看看新的表结构:
mysql> desc user; +-------+------------------+------+-----+---------+----------------+ | Field | Type | Null | Key | Default | Extra | +-------+------------------+------+-----+---------+----------------+ | id | int(10) unsigned | NO | PRI | NULL | auto_increment | | name | char(10) | NO | UNI | | | | extra | char(10) | NO | | | | +-------+------------------+------+-----+---------+----------------+
4、把数据导回去,在这里,有两种选择:新的重复记录替换旧的记录,只保留最新的记录
?或者是?新的记录略过,只保留最旧的记录
mysql> LOAD DATA INFILE '/tmp/user.txt' REPLACE INTO TABLE `user`; Query OK, 10 rows affected (0.00 sec) Records: 8 Deleted: 2 Skipped: 0 Warnings: 0 mysql> SELECT * FROM USER; +----+-------+--------+ | id | name | extra | +----+-------+--------+ | 1 | user1 | user1 | | 8 | user2 | user8 | | 6 | user3 | user6 | | 4 | user4 | user4 | | 5 | user5 | user5 | | 7 | user6 | user7 | | 9 | USER2 | user9 | | 10 | USER6 | user10 | +----+-------+--------+
?上面是采用?REPLACE
?的方式,可以看到,导入过程中删掉了两条数据,结果验证确实是?新的重复记录替换旧的记录,只保留最新的记录
。
?现在,来看看用?IGNORE
?的方式:
mysql> LOAD DATA INFILE '/tmp/user.txt' IGNORE INTO TABLE `user`; Query OK, 6 rows affected (0.01 sec) Records: 8 Deleted: 0 Skipped: 2 Warnings: 0 mysql> SELECT * FROM USER; +----+-------+--------+ | id | name | extra | +----+-------+--------+ | 1 | user1 | user1 | | 2 | user2 | user2 | | 3 | user3 | user3 | | 4 | user4 | us