SQL SERVER 的窗体函数OVER的使用：row_number/rank/dense_rank

举个例子给大家加深印象，也方便理解：

1.目前有这几笔数据：

Select '1班' banji,'张三' as name , as score
into #studentSoure
union all
Select '1班' banji,'李四' as name , as score
union all
Select '2班' banji,'王五' as name , as score
union all
Select '2班' banji,'王三' as name , as score
union all
Select '2班' banji,'王四' as name , as score

select * from #studentSoure 

2.简单明了看到所有同学和自己班级成绩最好的做比较

select banji,name,score,max(score) over(partition by banji) avg_score
from #studentSoure 

相同的道理，上面的max也可以改成min,avg等

3.row_number()、rank()、dense_rank()方法的使用

--row_number() 则在排序相同时不重复，会根据顺序排序。

--rank()排序相同时会重复，总数不会变 ，意思是会出现1、1、3这样的排序结果；

--dense_rank() 排序相同时会重复，总数会减少，意思是会出现1、1、2这样的排序结果。

select banji,name,score,ROW_NUMBER() over(partition by banji order by score) avg_score
from #studentSoure
select banji,name,score,rank() over(partition by banji order by score) avg_score
from #studentSoure
select banji,name,score,dense_rank() over(partition by banji order by score ) avg_score
from #studentSoure