partition  by关键字是分析性函数的一部分，它和聚合函数不同的地方在于它能返回一个分组中的多条记录，而聚合函数一般只有一条反映统计值的记录，partition  by用于给结果集分组，如果没有指定那么它把整个结果集作为一个分组

create database StudentDB
go

use StudentDB
go

create table Student  --学生成绩表
(
id int,  --主键
Grade int, --班级
Score int --分数
)
go

insert Student 
    select 1,1,88
union all select 2,1,66
union all select 3,1,75
union all select 4,2,30
union all select 5,2,70
union all select 6,2,80
union all select 7,2,60
union all select 8,3,90
union all select 9,3,70
union all select 10,3,80

go

--所有学生信息
select * from Student

id          Grade       Score
----------- ----------- -----------
1           1           88
2           1           66
3           1           75
4           2           30
5           2           70
6           2           80
7           2           60
8           3           90
9           3           70
10          3           80

(10 行受影响)

--不分班按学生成绩排名
select *,ROW_NUMBER() over(order by Score desc) as Sequence from Student

id          Grade       Score       Sequence
----------- ----------- ----------- --------------------
8           3           90          1
1           1           88          2
6           2           80          3
10          3           80          4
3           1           75          5
9           3           70          6
5           2           70          7
2           1           66          8
7           2           60          9
4           2           30          10

(10 行受影响)

--分班后按学生成绩排名
select *,row_number() over(partition by Grade order by Score desc) as Sequence from Student

id          Grade       Score       Sequence
----------- ----------- ----------- --------------------
1           1           88          1
3           1           75          2
2           1           66          3
6           2           80          1
5           2           70          2
7           2           60          3
4           2           30          4
8           3           90          1
10          3           80          2
9           3           70          3

(10 行受影响)