SQL中Group By的使用

select login_type,count(*) as count_sum,timestamp (DATE_FORMAT(create_time ,'%Y-%m-%d'),'%y-%m-%d') as login_time from sys_login_log where DATE_SUB(CURDATE(), INTERVAL 30 DAY) <=date(CREATE_TIME) GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d '),login_type;


select login_type,count(*) as count_sum,timestamp (DATE_FORMAT(create_time ,'%Y-%m-%d'),'%y-%m-%d') as login_time from sys_login_log where DATE_SUB(CURDATE(), INTERVAL 30 DAY) <=date(CREATE_TIME) and `user_id` IN (356,2399) GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d '),login_type;

SELECT login_type,sum(status) as count_sum,timestamp (DATE_FORMAT(create_time ,'%Y-%m-%d %H'),'%y-%m-%d %H') as login_time FROM sys_login_log WHERE DATE_FORMAT(create_time ,'%Y-%m-%d') LIKE concat('%', '2022-08-12', '%') GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d %H');

SELECT login_type,sum(status) as count_sum,timestamp (DATE_FORMAT(create_time ,'%Y-%m-%d %H'),'%y-%m-%d %H') as login_time FROM sys_login_log WHERE DATE_FORMAT(create_time ,'%Y-%m-%d') LIKE concat('%', '2022-08-12', '%') and `user_id` IN ('356,2399') GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d %H');


SELECT login_type,sum(status) as count_sum,timestamp (DATE_FORMAT(create_time ,'%Y-%m-%d %H'),'%y-%m-%d %H') as login_time FROM sys_login_log WHERE DATE_FORMAT(create_time ,'%Y-%m-%d') LIKE concat('%', '2022-08-12', '%') and FIND_IN_SET(user_account,'zyzhan1,wuchunyang') GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d %H');

按小时求和

SELECT sum(comment)as comment, SUM(payment_amount) as pay,DATE_FORMAT(begin_time ,'%Y-%m-%d %H') as begin_time FROM car_information GROUP BY DATE_FORMAT(begin_time ,'%Y-%m-%d %H')

• 按小时查询出了分组数据

按分类看出现的条数

SELECT  *,count(*) as com FROM images WHERE img_id = 52 GROUP BY des 

• 查询出了按DES 分组出来的数组和同名数量COUNT

2个条件的分组数据

SELECT  *,COUNT(des) as con FROM images,violatetable WHERE images.img_id = violatetable.car_id GROUP BY  img_id,des

• 查询出了联表按DES img_id分组的指定数组

多条件分组数组

SELECT  *,COUNT(des) as con,DATE_FORMAT(img_time ,'%Y-%m-%d %H') as imgtime FROM images,violatetable WHERE images.img_id = violatetable.car_id GROUP BY  img_id,des,DATE_FORMAT(img_time ,'%Y-%m-%d %H') ORDER BY imgtime

• 查询出了联表的按小时分组的指定数组

SQL中Group By的使用

• 1、概述
• 2、原始表
• 3、简单Group By
• 4、Group By 和 Order By
• 5、Group By中Select指定的字段限制
• 6、Group By All
• 7、Group By与聚合函数
• 8、Having与Where的区别
• 9、Compute 和 Compute By

1、概述

“Group By”从字面意义上理解就是根据“By”指定的规则对数据进行分组，所谓的分组就是将一个“数据集”划分成若干个“小区域”，然后针对若干个“小区域”进行数据处理。

2、原始表

3、简单Group By

示例1

select 类别, sum(数量) as 数量之和
from A
group by 类别

返回结果如下表，实际上就是分类汇总。

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-YWumbWME-1593330354684)(https://images0.cnblogs.com/blog/33509/201304/28234054-ff92ae14bfe74da98c4deb8d7c78f2f8.png)]

4、Group By 和 Order By

示例2

select 类别, sum(数量) AS 数量之和
from A
group by 类别
order by sum(数量) desc

返回结果如下表

在Access中不可以使用“order by 数量之和 desc”，但在SQL Server中则可以。

5、Group By中Select指定的字段限制

示例3

select 类别, sum(数量) as 数量之和, 摘要
from A
group by 类别
order by 类别 desc

示例3执行后会提示下错误，如下图。这就是需要注意的一点，在select指定的字段要么就要包含在Group By语句的后面，作为分组的依据；要么就要被包含在聚合函数中。

6、Group By All

示例4

select 类别, 摘要, sum(数量) as 数量之和
from A
group by all 类别, 摘要

示例4中则可以指定“摘要”字段，其原因在于“多列分组”中包含了“摘要字段”，其执行结果如下表

“多列分组”实际上就是就是按照多列（类别+摘要）合并后的值进行分组，示例4中可以看到“a, a2001, 13”为“a, a2001, 11”和“a, a2001, 2”两条记录的合并。

SQL Server中虽然支持“group by all”，但Microsoft SQL Server 的未来版本中将删除 GROUP BY ALL，避免在新的开发工作中使用 GROUP BY ALL。Access中是不支持“Group By All”的，但Access中同样支持多列分组，上述SQL Server中的SQL在Access可以写成

select 类别, 摘要, sum(数量) AS 数量之和
from A
group by 类别, 摘要

7、Group By与聚合函数

在示例3中提到group by语句中select指定的字段必须是“分组依据字段”，其他字段若想出现在select中则必须包含在聚合函数中，常见的聚合函数如下表：

函数	作用	支持性
sum(列名)	求和
max(列名)	最大值
min(列名)	最小值
avg(列名)	平均值
first(列名)	第一条记录	仅Access支持
last(列名)	最后一条记录	仅Access支持
count(列名)	统计记录数	注意和count(*)的区别

示例5：求各组平均值

select 类别, avg(数量) AS 平均值 from A group by 类别;

示例6：求各组记录数目

select 类别, count(*) AS 记录数 from A group by 类别;

示例7：求各组记录数目

8、Having与Where的区别

• where 子句的作用是在对查询结果进行分组前，将不符合where条件的行去掉，即在分组之前过滤数据，where条件中不能包含聚组函数，使用where条件过滤出特定的行。
• having 子句的作用是筛选满足条件的组，即在分组之后过滤数据，条件中经常包含聚组函数，使用having 条件过滤出特定的组，也可以使用多个分组标准进行分组。

示例8

select 类别, sum(数量) as 数量之和 from A
group by 类别
having sum(数量) > 18

示例9：Having和Where的联合使用方法

select 类别, SUM(数量)from A
where 数量 > 8
group by 类别
having SUM(数量) > 10

9、Compute 和 Compute By

select * from A where 数量 > 8

执行结果：

示例10：Compute

select *
from A
where 数量>8
compute max(数量),min(数量),avg(数量)

执行结果如下：

compute子句能够观察“查询结果”的数据细节或统计各列数据（如例10中max、min和avg），返回结果由select列表和compute统计结果组成。

示例11：Compute By

select *
from A
where 数量>8
order by 类别
compute max(数量),min(数量),avg(数量) by 类别

执行结果如下：

示例11与示例10相比多了“order by 类别”和“… by 类别”，示例10的执行结果实际是按照分组（a、b、c）进行了显示，每组都是由改组数据列表和改组数统计结果组成，另外：

• compute子句必须与order by子句用一起使用
• compute…by与group by相比，group by 只能得到各组数据的统计结果，而不能看到各组数据

在实际开发中compute与compute by的作用并不是很大，SQL Server支持compute和compute by，而Access并不支持

链接

实战代码保存

#　基本分组
SELECT login_type,count(*) as cou FROM sys_login_log GROUP BY login_type

# 当天
SELECT login_type,count(*) as cou  FROM sys_login_log WHERE DAY(create_time) = DAY(now()) GROUP BY login_type

# 昨天
SELECT login_type,count(*) as cou  FROM  sys_login_log WHERE TO_DAYS( NOW( ) ) - TO_DAYS( create_time) = 1  GROUP BY login_type

# 近7天  
SELECT login_type,count(*) as cou  FROM  sys_login_log where DATE_SUB(CURDATE(), INTERVAL 7 DAY) <= date(create_time)  GROUP BY login_type

# 近30天  
SELECT login_type,count(*) as cou  FROM sys_login_log where DATE_SUB(CURDATE(), INTERVAL 30 DAY) <= date(create_time) GROUP BY login_type

#本周
SELECT login_type,count(*) as cou  FROM sys_login_log WHERE WEEK(create_time) = WEEK(now()) GROUP BY login_type
#本月
SELECT login_type,count(*) as cou  FROM sys_login_log WHERE MONTH(create_time) = MONTH(now()) GROUP BY login_type
#本季
select login_type,count(*) as cou  from `sys_login_log` where QUARTER(create_time)=QUARTER(now())  GROUP BY login_type 
#上一月  
SELECT login_type,count(*)  FROM sys_login_log WHERE PERIOD_DIFF( date_format( now( ) , '%Y%m' ) , date_format( create_time, '%Y%m' ) ) =1   GROUP BY login_type;  
# 查询距离当前现在6个月的数据  
select login_type,count(*)  from sys_login_log where create_time between date_sub(now(),interval 6 month) and now()  GROUP BY login_type
# 当年
SELECT login_type,count(*) as cou  FROM sys_login_log WHERE YEAR(create_time) = YEAR(now()) GROUP BY login_type
SELECT login_type,count(*) as CountSum  FROM sys_login_log WHERE YEAR(create_time) = '2020' GROUP BY login_type
#查询上年数据  
select login_type,count(*) from `sys_login_log` where year(create_time)=year(date_sub(now(),interval 1 year))  GROUP BY login_type

X年X月
SELECT login_type,count(*) as count_sum  FROM sys_login_log WHERE YEAR(create_time)='2020' and MONTH(create_time)='12' GROUP BY login_type



#　按天分
SELECT login_type,sum(status) as con,DATE_FORMAT(create_time ,'%Y-%m-%d ') as login_time FROM sys_login_log where create_time between date_sub(now(),interval 1 MONTH) and now()   GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d '),login_type

#　当天按小时
SELECT sum(status) as con,DATE_FORMAT(create_time ,'%Y-%m-%d %H') as login_time FROM sys_login_log WHERE DATE_FORMAT(create_time ,'%Y-%m-%d') LIKE '%2020-12-12%' GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d %H')

#　时间范围按小时
SELECT sum(status) as con,DATE_FORMAT(create_time ,'%Y-%m-%d %H') as login_time FROM sys_login_log WHERE create_time BETWEEN '2020-12-8' AND '2020-12-12'  GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d %H')

#　指定代理的用户登陆统计按天
SELECT sum(status) as con,DATE_FORMAT(create_time ,'%Y-%m-%d ') as login_time FROM sys_login_log WHERE user_id BETWEEN 400 AND 450 GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d ')

SELECT login_type,sum(status) as con,DATE_FORMAT(create_time ,'%Y-%m-%d ') as login_time FROM sys_login_log WHERE user_id in (400,401,402,403,404,405) GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d '),login_type

 
# 增加代理条件
SELECT login_type,sum(status) as con,DATE_FORMAT(create_time ,'%Y-%m-%d ') as login_time FROM sys_login_log WHERE user_id IN (400,401,402)  GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d '),login_type

# 查询代理的数据
SELECT * FROM sys_login_log WHERE user_id IN (400,401,402)


#　指定代理的用户登陆统计按天
    SELECT a.login_type,sum(a.status) as count_sum,a.create_time FROM sys_login_log  a WHERE a.user_id  in
	 (SELECT user_id from sys_user  WHERE adder_id = 278)  GROUP BY DATE_FORMAT(create_time ,'%Y-%m-%d '),login_type

 SELECT a.login_type,sum(a.status) as count_sum FROM sys_login_log  a WHERE a.user_id  in
	 (SELECT user_id from sys_user  WHERE adder_id = 278)  GROUP BY login_type