当前位置：首页 >数据库 > MySql 语句收集

MySql 语句收集

=与:=区别

= 只有在set和update时才是和:=一样，赋值的作用，其它都是等于的作用。鉴于此，用变量实现行号时，必须用:=
:= 不只在set和update时时赋值的作用，在select也是赋值的作用。

序列号：

两种创建变量并赋值的方式
SET @变量名 = 值;
SELECT 值 INTO @变量名;

select (@rowNO := @rowNo+1) AS rowno from blog,(select @rowNO :=0) b

必须要有select @rowNo:=0,不然查询出来的全是null

或者这样

set @rowNo=0;select (@rowNO := @rowNo+1) AS rowno from blog;

分组：

一个user表，有id,name字段，name有重复，求id最小的不同name的记录表。

复杂的：

select * from user a where a.id=(select id from user b where a.name=b.name limit 1)

group by：

select * from user where id in (select  min(id) from user group by name having count(1)>0)

使用

select ta.* from hs_workflow_log as ta,(select max(process_instance_id) as process_instance_id,business_no from hs_workflow_log  group by business_no) as tbwhere ta.business_no=tb.business_no and ta.process_instance_id=tb.process_instance_idand  ta.business_no in('FD188189405673029642', 'FD188189208538644488', 'FD188188780774162444')order by ta.create_time desc

子查询分组：

select * from (select * from `test` order by `date` desc) `temp` group by category_id order by `date` desc

虽然没有理解，但是测下来是对的

上面三种，测试下来，还是groupby与子查询分组效率差不多，最上面的效率最低，不管是否有索引的情况，都最慢。

上面的方式mysql5.7以后无效，修改为：

select * from (select * from `test` order by `date` desc limit 10000) `temp` group by category_id order by `date` desc

可能是因为子查询大多数是作为一个结果给主查询使用，所以子查询不需要排序的原因。

对子查询的排序进行limit限制，此时子查询就不光是排序，所以此时排序会生效，但是限制条数却只能尽可能的设置大些。

分组数量查询：

现在希望取整个表中business_no中的process_instance_id不同的数量

select business_no,count(distinct process_instance_id) as num from hs_workflow_log  group by business_no having num>1

分组，组合拼接字段

CREATE TABLE `USERS`(`Id` INT(4) NOT NULL,`Name` VARCHAR(30) DEFAULT '');INSERT INTO USERS(`Id`,`Name`) VALUES(1,'A'),(1,'B'),(2,'C');SELECT Id,GROUP_CONCAT(`Name` SEPARATOR ',') NAMES FROM `USERS` GROUP BY Id;#不使用分隔符SELECT Id,GROUP_CONCAT(Name order by Name) NAMES FROM `USERS` GROUP BY Id;

GROUP_CONCAT()函数的分隔符可以自定义，根据你的需求适当选择即可

用group_concat的时候请注意，连接起来的字段如果是int型，一定要转换成char再拼起来,否则返回的是byte[]

该问题当你在SQLyog等一些工具中是体现不出来的，所以很难发现。select group_concat(ipaddress) from t_ip 返回逗号隔开的串select group_concat(id) from t_ip 返回byte[]select group_concat(CAST(id as char)) from t_dep 返回逗号隔开的串select group_concat(Convert(id , char)) from t_dep 返回逗号隔开的串

group_concat连接字段的时候是有长度限制的，并不是有多少连多少。但可以设置一下

同数据库表数据迁移

-- 迁移统一账户数据update hs_issue_info a ,hs_draw_info b set a.add_person_p2p_account_draw=1,a.add_person_p2p_account_draw_archive_id=b.unify_loan_account,a.add_person_p2p_account_draw_p2p_account_id=b.p2p_account_id,a.add_person_p2p_account_draw_p2p_account_name=b.p2p_account_name,a.add_person_p2p_account_draw_p2p_account_tel=b.phone_numberwhere a.draw_id=b.draw_id and b.unify_loan_account is not null;

存储过程

delimiter $$drop procedure if exists pro;create procedure pro()begindeclare num int default 0;	 select count(1) from hs_draw_info where unify_loan_account is not null into num;	 select num;end $$delimiter ;call pro();

delimiter 是分隔符，其实就是告诉mysql解释器，该段命令是否已经结束了，mysql是否可以执行了。
默认情况下，delimiter是分号;，在命令行客户端中，如果有一行命令以分号结束，那么回车后，mysql将会执行该命令。

最后一个$$就是告诉命令行，语句可以执行了。

案例，迁移数据

创建临时表，遍历插入数据

-- 创建临时表CREATE TEMPORARY TABLE IF NOT EXISTS temp(	id INT NOT NULL DEFAULT 0,	role_full_name VARCHAR(100) NOT NULL,	right_full_name VARCHAR(100) NOT NULL)-- 初始化数据set @rowNo=0;insert into temp (id,role_full_name,right_full_name) select  (@rowNO := @rowNo+1) AS id,a.full_name as role_full_name,c.full_name as right_full_name  from role a join role_right b on a.role_id=b.role_idjoin `right` con b.right_id=c.right_idwhere a.application_id=39and c.application_id=39;-- 修改数据-- 修改数据update temp set role_full_name=replace(role_full_name,'huishi','huishi-pub');update temp set right_full_name=replace(right_full_name,'huishi','huishi-pub');update temp set role_full_name=replace(role_full_name,'huishi','huishi-pub');update temp set right_full_name=replace(right_full_name,'huishi','huishi-pub');-- 查看信息select * from temp;select distinct role_full_name,right_full_name from temp;select @rowNo;-- 创建存储过程，循环插入数据delimiter $$drop procedure if exists pro $$create procedure pro()begindeclare num int default 1;	 declare totalNum int default 0;	 declare role_id2 int default 0;	 declare right_id2 int default 0;declare role_full_name2 varchar(100) default '';	 declare right_full_name2 varchar(100) default '';	 set num=1;	 while num<=@rowNo do			 select role_full_name,right_full_name into role_full_name2,right_full_name2  from temp where id=num;			 select role_id into role_id2 from role where full_name=role_full_name2;			 select right_id into right_id2 from `right` where full_name=right_full_name2;			 select count(1) into totalNum from role_right where role_id=role_id2 and right_id=right_id2;			 if totalNum=0 then			  INSERT INTO `role_right`(`role_id`, `right_id`, `creator`, `is_active`, `insert_time`, `update_time`) VALUES (role_id2, right_id2, NULL, 1, now(), now());			 end if;			 set num=num+1;	 end while;end $$delimiter ;-- 执行存储过程call pro();-- 删除临时表DROP TABLE IF EXISTS temp;

迁移部分征信数据附件：

delimiter $$drop procedure if exists pro $$create procedure pro()begindeclare num,dataNum int default 0;declare archiveId,searchArchiveId,creditDataId bigint default 0;set num=1;  set @rowNo=0;create table temp asselect  (@rowNO := @rowNo+1) AS id, a.*from hs_credit_order ajoin hs_loan_order bon a.loan_order_id=b.loan_order_idjoin hs_workflow_log con a.loan_order_id=c.business_nowhere b.loan_status=7and c.workflow_key like 'HS-FKSP%'and c.current_node_name='通过'and c.create_time>'2020-02-18 00:00:00'and c.create_time<'2020-05-19 00:00:00';	 while num<=@rowNo do			 select dataNum=count(1) from hs_cdc_credit_data a join temp b on a.credit_order_id=b.credit_order_id;			 -- 只有不存在对应征信才新增数据			 if dataNum=0 then			INSERT INTO  `hs_cdc_credit_data`(  `loan_order_id`, `credit_order_id`, `credit_category`, `business_category`, `search_name`, `search_archive_id`, `report_time`, `complete_time`, `credit_data_status`, `create_time`, `create_name`, `modify_time`, `modify_name`) 				 select loan_order_id,credit_order_id,2,2,search_name,search_archive_id,NULL,NULL,1,now(),'admin',now(),'admin' from temp where id=num;				 select search_archive_id into searchArchiveId from temp where id=num;				 select LAST_INSERT_ID() into creditDataId;				 -- 批量新增查询附件				 INSERT INTO `hs_cdc_credit_data_file`( `credit_data_id`, `catalog_key`, `catalog_name`, `image_id`, `file_sort`, `is_image`, `is_system`, `file_status`, `create_time`, `create_name`, `modify_time`, `modify_name`)select creditDataId,NULL,NULL,bb.img_id,bb.detail_sort,true,true,1,now(),'admin',now(),'admin' from hs_archive_attachment aa joinhs_archive_attachment_detail bb on aa.attachment_id=bb.attachment_idwhere aa.catalog_key='rhzxbgtp' and aa.attachment_status=1 and bb.detail_status=1 and aa.archive_id=searchArchiveId;				 	 			 end if;			 set num=num+1;	 end while;	 	 drop table temp;end $$delimiter ;-- 执行存储过程call pro();

注意上面，创建临时表使用了