partition by

本文转载，地址：https://blog.csdn.net/dwt1415403329/article/details/87835383

1、建表并插入数据

因为over是开窗函数，mysql不支持开窗函数，其他如oracle,sql server,db2…等新版本都支持

CREATE TABLE Employee
(
  ID number(10) not null primary key,
  EmpName varchar(20),
  EmpSalary varchar(10),
  EmpDepartment varchar(20)
);
 
insert all into Employee values(1,'张三','5000','开发部')
into Employee values(2,'李四','2000','销售部')
into Employee values(3,'王麻子','2500','销售部')
into Employee values(4,'张三表叔','8000','开发部')
into Employee values(5,'李四表叔','5000','开发部')
into Employee values(6,'王麻子表叔','5000','销售部')
select 1 from dual

2、计算每个部门的总薪水

2.1、group by

SELECT EmpDepartment,SUM(EmpSalary) sum_sala FROM Employee GROUP BY EmpDepartment

2.2、over(partition by)

SELECT EmpSalary,EmpDepartment,SUM(EmpSalary) OVER(PARTITION BY EmpDepartment) sum_sala FROM Employee

小结：
group by 和 partition by 都有分组统计的功能，但partition by并没有group by的汇总功能。
partition by统计的每一条记录都存在，而group by将所有的记录汇总成一条记录(类似于distinct EmpDepartment 去重)。
partition by可以和聚合函数结合使用，同时具有其他高级功能。

3、partition by继续探究

SELECT EmpSalary,EmpDepartment,SUM(EmpSalary) OVER(PARTITION BY EmpDepartment ORDER BY EmpSalary) sum_sala FROM Employee

小结：

加上order by 后，类似于累加功能(sum_sala = sum_sala + EmpSalary)
先观察销售部的结果: 
	第4条记录: sum(EmpSalary)即sum_sala=2000，
	第5条记录: sum(EmpSalary)=sum_sala+2500=4500，
	即第4条sum_sala与第5条EmpSalary的和，依次类推；
开发部：
	由于2个5000是并列的，所以计算的时候是几个并列数据之和即5000+5000=10000。

with t1 as (select * from (
  select a.group_id,a.ap_start_date ,a.row_id, row_number() over(partition by t.group_Id order by t.TASK_DATE desc) mm
  from em_task t
       left join Em_Re_Group a on a.task_id = t.row_id where t.task_Type in( '1' ,'4') and t.status_Flag = '9'
       and t.c_Status_Flag = '4' and t.deleted_flag = '0') tt where mm = 1)

select q.dealer_code,t1.group_id,q.dealer_name,t1.ap_start_date from dm_dealer q, t1  where q.group_id = t1.group_id;