手机版
你好,游客 登录 注册 搜索
背景:
阅读新闻

MySQL触发器案例分析及before与after的区别

[日期:2018-04-26] 来源:Linux社区  作者:贺子 [字体: ]

触发器(trigger):监视某种情况,并触发某种操作,它是提供给程序员和数据分析员来保证数据完整性的一种方法,它是与表事件相关的特殊的存储过程,它的执行不是由程序调用,也不是手工启动,而是由事件来触发,例如当对一个表进行操作( insert,delete, update)时就会激活它执行。

触发器经常用于加强数据的完整性约束和业务规则等。 触发器创建语法四要素:

1.监视地点(table)

2.监视事件(insert/update/delete)

3.触发时间(after/before)

4.触发事件(insert/update/delete)

其中:trigger_time是触发器的触发事件,可以为before(在检查约束前触发)或after(在检查约束后触发);trigger_event是触发器的触发事件,包括insert、update和delete,可以使用old和new来引用触发器中发生变化的记录内容。

需要注意的:

1)需注意对同一个表的相同触发时间(after/before)的相同触发事件(insert/update/delete),只能定义一个触发器,否则报错

ERROR 1235 (42000): This version of MySQL doesn't yet support 'multiple triggers with the same action time and event for one table'

2)mysql触发器中的if判断语法格式如下:(1).循环中还可以有循环,(2).else后面没有then, (3).elseif!!!!不是else if !!

if...then{

if...then{}

end if;

if...then{}

end if;

...

}

elseif...then..

else

end if ;

注意可以使用两个if循环,各自end if即可

3)注意mysql触发器中的before和after的区别:

before:(insert、update)可以对new进行修改,

after:不能对new进行修改,两者都不能修改old数据。

对于INSERT语句, 只有NEW是合法的;

对于DELETE语句,只有OLD才合法;

对于UPDATE语句,NEW、OLD可以同时使用。

after是先完成数据的增删改,再触发,触发的语句晚于监视的增删改操作,无法影响前面的增删改动作;也就是说先插入订单记录,再更新商品的数量;

before是先完成触发,再增删改,触发的语句先于监视的增删改,这样就可以对new进行修改了;

摘自网络的一个例子说明:

首先我们来创建两张表:

#商品表

create table g

(

  id int primary key auto_increment,

  name varchar(20),

  num int

);

#订单表

create table o

(

  oid int primary key auto_increment,

  gid int,

much int

);

insert into g(name,num) values('商品1',10),('商品2',10),('商品3',10);

我们借助触发器来完成下订单之后,自动对商品表中相应商品做减法;如下:

create trigger tg2

after insert on o

for each row

begin

update g set num=num-new.much where id=new.gid;

end$

但是有个问题是,如果下订单数超过商品总数时,那么会导致商品表中产生负数,这样我们可以借助before来对订单中new值进行修改,保证商品表不会出现负数;

案例:当新增一条订单记录时,判断订单的商品数量,如果数量大于10,就默认改为10

DELIMITER $

create trigger tg6

beforeinsert on o

for each row

begin

  if new.much > 10 then

    set new.much = 10;

  end if;

  update g set num = num - new.much where id = new.gid;

end $

DELIMITER ;

4)不是说一个事务出发一次,如下这个事务修改了10行数,他会触发10次:

mysql> update blocks_infos set infos_id=1 where infos_id=2;

Query OK, 10 rows affected (0.22 sec)

Rows matched: 10 Changed: 10 Warnings: 0

5)针对before的情况,如果触发的操作没有成功,会导致原本的触发事件也不成功;

接下来记录下,我写的案例,当对一个表做增删改的时候,触发对另一表做相应的操作,

例如下面,如果begin后面有语法错误或者执行错误,那么会导致前面的delete失败;

DELIMITER $

create trigger tri_delete_blocks_infos1 before delete

on blocks_infos for each row

begin

DECLARE h int;

set h=(select intc from bidinfo.v_publish_info where id=old.infos_id);

if h is null then

update bidinfo.v_publish_info set intc=1 where id= old.infos_id;

else

update bidinfo.v_publish_info set intc=intc+1 where id= old.infos_id;

end if;

end $

DELIMITER ;

1.关于insert的触发器:

我们的要求是当向blocks_infos的时候,先判断blocks_infos_opensearch表中有没有新insert的infos_id,如果有就相应的update,没有的话就insert,可以如下两种方法:

方法一使用replace:

DELIMITER $

create trigger tri_insert_blocks_infos after insert

on blocks_infos for each row

begin

replace into blocks_infos_opensearch (infos_id,blocks) select infos_id,group_concat(blocks_id) blocks from blocks_infos where infos_id=new.infos_id group by infos_id;

end $

DELIMITER ;

注意关于MySQL replace into 有三种形式(into关键字可以省略):

1. replace into tbl_name(col_name, ...) values(...)

2. replace into tbl_name(col_name, ...) select ...

3. replace into tbl_name set col_name=value, ...

方法二:用if判断:

DELIMITER $

create trigger tri_insert_blocks_infos after insert

on blocks_infos for each row

begin

DECLARE c INT;

set c=(SELECT COUNT(infos_id) FROM blocks_infos WHERE infos_id=new.infos_id);

if c=1 then

insert into blocks_infos_opensearch select infos_id,GROUP_CONCAT(blocks_id) blocks FROM blocks_infos WHERE infos_id=new.infos_id;

elseif c>1 then

UPDATE blocks_infos_opensearch SET blocks= (SELECT GROUP_CONCAT(blocks_id) blocks FROM blocks_infos WHERE infos_id=new.infos_id ) WHERE infos_id= new.infos_id;

end if ;

end $

DELIMITER ;

2.关于delete的触发器:

DELIMITER $

CREATE TRIGGER tri_delete_blocks_infos after DELETE

ON blocks_infos FOR EACH ROW

BEGIN

DECLARE c INT;

SET c=(SELECT COUNT(infos_id) FROM blocks_infos WHERE infos_id=old.infos_id);

IF c=0 THEN

DELETE FROM blocks_infos_opensearch WHERE infos_id=old.infos_id;

ELSEIF c>0 THEN

UPDATE blocks_infos_opensearch SET blocks= (SELECT GROUP_CONCAT(blocks_id) blocks FROM blocks_infos WHERE infos_id=old.infos_id ) WHERE infos_id= old.infos_id;

END IF;

END $

DELIMITER ;

3.关于update的触发器:

DELIMITER $

CREATE TRIGGER tri_update_blocks_infos after update

ON blocks_infos FOR EACH ROW

BEGIN

DECLARE c INT;

DECLARE d varchar(1000);

DECLARE h varchar(1000);

SET c=(SELECT COUNT(infos_id) FROM blocks_infos WHERE infos_id=old.infos_id);

set d=(SELECT GROUP_CONCAT(blocks_id) blocks FROM blocks_infos WHERE infos_id=old.infos_id);

set h=(SELECT GROUP_CONCAT(blocks_id) blocks FROM blocks_infos WHERE infos_id=new.infos_id);

IF c=0 THEN

DELETE FROM blocks_infos_opensearch WHERE infos_id=old.infos_id;

ELSEIF c>0 THEN

UPDATE blocks_infos_opensearch SET blocks= d WHERE infos_id= old.infos_id;

UPDATE blocks_infos_opensearch SET blocks= h WHERE infos_id= new.infos_id;

END IF;

END $

DELIMITER ;

另一个需求是需要当对表blocks_infos做相关处理的时候,会触发另一个表bidinfo.v_publish_info 做相应的处理,因为前面已经建立了after insert on blocks_infos,不能再建立 after insert

on blocks_infos,所以只能创建 before insert on blocks_infos,如下创建了三个:

1)insert

DELIMITER $

create trigger tri_insert_blocks_infos1 before insert

on blocks_infos for each row

begin

DECLARE d int;

set d=(select intc from bidinfo.v_publish_info where id=new.infos_id);

if d is null then

update bidinfo.v_publish_info set intc=1 where id= new.infos_id;

else

update bidinfo.v_publish_info set intc=intc+1 where id= new.infos_id;

end if;

end $

DELIMITER ;

2)delete

DELIMITER $

create trigger tri_delete_blocks_infos1 before delete

on blocks_infos for each row

begin

DECLARE h int;

set h=(select intc from bidinfo.v_publish_info where id=old.infos_id);

if h is null then

update bidinfo.v_publish_info set intc=1 where id= old.infos_id;

else

update bidinfo.v_publish_info set intc=intc+1 where id= old.infos_id;

end if;

end $

DELIMITER ;

3)update ,注意可以只用两个if循环!

DELIMITER $

create trigger tri_update_blocks_infos1 before update

on blocks_infos for each row

begin

DECLARE j int;

DECLARE i int;

set i=(select intc from bidinfo.v_publish_info where id=new.infos_id);

set j=(select intc from bidinfo.v_publish_info where id=old.infos_id);

if j is null then

update bidinfo.v_publish_info set intc=1 where id= old.infos_id;

else

update bidinfo.v_publish_info set intc=intc+1 where id= old.infos_id;

end if;

if i is null then

update bidinfo.v_publish_info set intc=1 where id= new.infos_id;

else

update bidinfo.v_publish_info set intc=intc+1 where id= new.infos_id;

end if;

end $

DELIMITER ;

小结:触发器中的new和old,可以理解为处理过的整行数据,可以通过new.字段名来取出那个字段的值,并且alter和before都不能修改old的值,但是before可以修改new的值,还需要注意对同一个表的相同触发时间(after/before)的相同触发事件(insert/update/delete),只能定义一个触发器,并且before的形式的触发器需要保证触发端和被触发端都得成功才能成功!

本文永久更新链接地址https://www.linuxidc.com/Linux/2018-04/152079.htm

linux
相关资讯       MySQL触发器 
本文评论   查看全部评论 (0)
表情: 表情 姓名: 字数

       

评论声明
  • 尊重网上道德,遵守中华人民共和国的各项有关法律法规
  • 承担一切因您的行为而直接或间接导致的民事或刑事法律责任
  • 本站管理人员有权保留或删除其管辖留言中的任意内容
  • 本站有权在网站内转载或引用您的评论
  • 参与本评论即表明您已经阅读并接受上述条款