内容简介:MySQL限制数据重复的方式:表上增加主键(Primary Key)或增加唯一性索引(Unique)主键对重复资料进行限制,这样资料在导入时就无法重复插入唯一性索引对复制资料进行限制
MySQL限制数据重复的方式:表上增加主键(Primary Key)或增加唯一性索引(Unique)
主键对重复资料进行限制,这样资料在导入时就无法重复插入
create table primary_t( last_name varchar(20), first_name varchar(20), sex varchar(8), primary key(last_name,first_name) )engine=innodb default charset=utf8; show table status like 'primary_t'\G show index from primary_t\G 模拟插入两笔有主键约束的资料 insert into primary_t(last_name,first_name,sex) values('Frank','Li','Man'),('Frank','Li','Man'),('Coco','Li','Woman'); ERROR 1062 (23000): Duplicate entry 'Frank-Li' for key 'PRIMARY' insert into primary_t(last_name,first_name,sex) values('Frank','Li','Man'),('Frank','Cai','Man'),('Coco','Li','Woman'); mysql> select * from primary_t; +-----------+------------+-------+ | last_name | first_name | sex | +-----------+------------+-------+ | Coco | Li | Woman | | Frank | Cai | Man | | Frank | Li | Man | +-----------+------------+-------+ rows in set (0.00 sec)
唯一性索引对复制资料进行限制
create table unique_t( last_name varchar(20), first_name varchar(20), sex varchar(8), unique index unique_name(last_name,first_name) )engine=innodb default charset=utf8; show table status like 'unique_t'\G show index from unique_t\G 模拟插入两笔有唯一性约束的资料 insert into unique_t(last_name,first_name,sex) values('Frank','Li','Man'),('Frank','Li','Man'),('Coco','Li','Woman'); ERROR 1062 (23000): Duplicate entry 'Frank-Li' for key 'unique_name' insert into unique_t(last_name,first_name,sex) values('Frank','Li','Man'),('Frank','Cai','Man'),('Coco','Li','Woman'); mysql> select * from unique_t; +-----------+------------+-------+ | last_name | first_name | sex | +-----------+------------+-------+ | Frank | Li | Man | | Frank | Cai | Man | | Coco | Li | Woman | +-----------+------------+-------+ rows in set (0.00 sec)
对资料的统计
create table count_t( last_name varchar(20), first_name varchar(20), sex varchar(8) )engine=innodb default charset=utf8; insert into count_t values('Frank','A','Man'), ('Frank','B','Man'), ('Frank','C','Woman'), ('Frank','D','Man'), ('Frank','E','Man'), ('Frank','F','Woman'), ('Frank','G','Man'), ('Frank','H','Man'), ('Frank','I','Woman'), ('Coco','A','Woman'), ('Coco','B','Man'), ('Coco','C','Man'), ('Coco','D','Man'), ('Coco','E','Man'), ('Coco','F','Man'), ('Coco','G','Woman'), ('Coco','H','Woman'), ('Alex','B','Man'), ('ALex','C','Man'), ('ALex','D','Man'), ('Alex','E','Man'), ('ALex','F','Man'), ('ALex','G','Woman'), ('ALex','H','Woman') ; commit; 按各个列进行分组统计 select last_name,count(*) from count_t group by last_name; select first_name,count(*) from count_t group by first_name; select sex,count(*) from count_t group by sex; mysql> select last_name,count(*) from count_t group by last_name; +-----------+----------+ | last_name | count(*) | +-----------+----------+ | Alex | 7 | | Coco | 8 | | Frank | 9 | +-----------+----------+ rows in set (0.00 sec) mysql> select first_name,count(*) from count_t group by first_name; +------------+----------+ | first_name | count(*) | +------------+----------+ | A | 2 | | B | 3 | | C | 3 | | D | 3 | | E | 3 | | F | 3 | | G | 3 | | H | 3 | | I | 1 | +------------+----------+ rows in set (0.00 sec) mysql> select sex,count(*) from count_t group by sex; +-------+----------+ | sex | count(*) | +-------+----------+ | Man | 16 | | Woman | 8 | +-------+----------+ rows in set (0.00 sec) 对分组的统计在进行限定,having 大于等于8次的记录 select last_name ,count(*) from count_t group by last_name having count(*) >=8; mysql> select last_name ,count(*) from count_t group by last_name having count(*) >=8; +-----------+----------+ | last_name | count(*) | +-----------+----------+ | Coco | 8 | | Frank | 9 | +-----------+----------+ rows in set (0.00 sec)
对资料进行去重操作,主要使用distinct函数
select distinct last_name from count_t; select distinct first_name from count_t; select distinct sex from count_t;
对已有资料的表进行去重操作
insert into count_t values('Xi','Xi','Man'),('Xi','Xi','Man'); 增加主键约束,加入ignore,忽悠重复的资料 alter ignore table count_t add primary key(last_name,first_name); select * from count_t where last_name='Xi'; mysql> alter ignore table count_t -> add primary key(last_name,first_name); Query OK, 26 rows affected (0.01 sec) Records: 26 Duplicates: 1 Warnings: 0 mysql> select * from count_t where last_name='Xi'; +-----------+------------+------+ | last_name | first_name | sex | +-----------+------------+------+ | Xi | Xi | Man | +-----------+------------+------+ row in set (0.00 sec) 使用group by方式也可去除重复资料 create table count_t0 as select last_name,first_name,sex from count_t group by last_name,first_name,sex;
Linux公社的RSS地址 : https://www.linuxidc.com/rssFeed.aspx
本文永久更新链接地址: https://www.linuxidc.com/Linux/2019-05/158748.htm
以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持 码农网
猜你喜欢:本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们。
移动交互设计精髓
班格 (Cameron Banga)、温霍尔德 (Josh Weinhold) / 傅小贞、张颖鋆 / 电子工业出版社 / 2015-4-1 / CNY 89.00
越来越多的人正涌入移动应用领域,而设计和体验将是移动应用成败的关键。作者通过上百款应用的设计实践,系统化地梳理了移动应用的设计方法论,在理解用户、跨平台和适配设计、移动组件应用、界面视觉感染力、简约设计等方面都进行了深入阐述。此外,作者还介绍了一些非常实用的移动设计工具,分享了设计师该如何与开发工程师协同工作,以及如何收集用户反馈、甄别版本迭代的更新需求等。 《移动交互设计精髓——设计完美的......一起来看看 《移动交互设计精髓》 这本书的介绍吧!