Sql server 开窗函数over()的语法

用法一：与ROW_NUMBER()函数结合用，给结果进行排序编号，如图：

代码如下：

SELECT ROW_NUMBER() over(order by RequiredDate) num
       ,* from  [Northwind].[dbo].[Orders]

用法二：跟聚合函数一起使用，利用over子句的分组效率比group by子句的效率更高。

在Northwind数据库的订单表Orders中查询"订单id","客户id","运费","所有订单的总数"，“每一个客户的总运费”，“所有客户的总运费”，“每一个客户的平均运费”，“所有客户的平均运费”,"每一个客户所有订单中最大的运费","所有客户中最大运费","每一个客户所有订单中最小的运费","所有客户中最小运费",如下图:

代码如下：

  SELECT  [OrderID] --订单id
      ,[CustomerID] --客户id
      ,[Freight] --运费
      ,COUNT(OrderID) over() as totalNum --一所有订单的总数
      ,SUM(Freight) over(partition by customerid) as cusTotalFreight --每一个客户的总运费
      ,SUM(Freight) over() as totalFreight --所有客户的总运费
      ,AVG(Freight) over(partition by customerid) as cusAvgFreight --每一个客户的平均运费
      ,AVG(Freight) over() as avgFreight --所有客户的平均运费
      ,MAX(Freight) over(partition by customerid) as cusMaxFreight --每一个客户所有订单中最大的运费
      ,MAX(Freight) over() as maxFreight --所有客户中最大运费
      ,MIN(Freight) over(partition by customerid) as cusMinFreight --每一个客户所有订单中最小的运费
      ,MIN(Freight) over() as minFreight --所有客户中最小运费
  FROM [Northwind].[dbo].[Orders]

总结：运用开窗函数over()再结合聚合函数，可以使表格向右扩张，并进行一些数据的统计