T-SQL基础(2)

开窗函数over

select orderid, custid, val,
SUM(val) over() as totalvalue,
SUM(val) over(partition by custid) as custtotalvalue
from Sales.OrderValues

select orderid, custid, val,
100. * val / SUM(val) over() as pctall,
100. * val / SUM(val) over(partition by custid) as pctcust
from Sales.OrderValues

over 子句也支持四种排名函数：row_number, rank, dense_rank and ntile

select orderid, custid, val,
　　row_number() over(order by val) as rownum,
　　rank() over(order by val) as rank,
　　dense_rank() over(order by val) as dense_rank,
　　ntile(10) over(order by val) as ntile --分为10组，显示当前行所在的组号
from Sales.OrderValues
order by val;

select orderid, custid, val,
row_number() over(partition by custid order by val) as rownum
from Sales.OrderValues
order by custid, val;

如果在select处理阶段指定了开窗函数，开窗函数必须在distinct子句(如果存在)之前进行处理。
OrderValues视图目前有830行，795个不同的val值。

select distinct val, ROW_NUMBER() over(order by val) as rownum
from Sales.OrderValues
--返回830 rows

select distinct val from Sales.OrderValues
--返回795 rows

可以认为在同一select子句中不同时指定distinct和row_number是一条最佳实践原则，因为distinct子句在这种情况下不起任何作用。可以使用group by为795个唯一值分配行号。

select val, ROW_NUMBER() over(order by val) as rownum
from Sales.OrderValues group by val;

select 各子句处理顺序

from
where
group by
having
select
　　over 
　　distinct
　　top
order by