SQL排序之 row_number, rank(), dense_rank()区别

三者的区别如下：

rank()排序相同时会重复，总数不变，即会出现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，即会出现1、1、2这样的排序结果；
row_number()排序相同时不会重复，会根据顺序排序。

具体实例

建表、插入数据

create table rownumber(
       id varchar(10) not null,
       name varchar(10) null,
       age varchar(10) null,
       salary int null
);

insert into rownumber(id,name,age,salary) values(1,'a',10,8000);
insert into rownumber(id,name,age,salary) values(1,'a2',11,7500);
insert into rownumber(id,name,age,salary) values(2,'b',12,7500);
insert into rownumber(id,name,age,salary) values(2,'b2',13,4500);
insert into rownumber(id,name,age,salary) values(3,'c',14,8000);
insert into rownumber(id,name,age,salary) values(3,'c2',15,20000);
insert into rownumber(id,name,age,salary) values(4,'d',16,30000);
insert into rownumber(id,name,age,salary) values(5,'d2',17,8000);
select * from rownumber;

数据

row_number()

select *, row_number() over(order by salary) as `rank` from rownumber;

row_number()排序结果

rank()

select *, rank() over(order by salary) as `rank` from rownumber;

rank()排序结果

dense_rank()

select *, dense_rank() over(order by salary) as `rank` from rownumber;

dense_rank()排序结果

row_number()的详细用法

1.分组排序

select *, row_number() over(partition by id order by salary) as `rank` from rownumber;

2. 每组内第二名的信息

select * from (
select *, row_number() over(partition by id order by salary) as `rank` from 
rownumber) as temp
where `rank` = 2;

3. 给年龄在13岁到16岁的数据按salary排名

select *, row_number() over (order by salary) as `rank` from rownumber
where age between 13 and 16;

注：

row_number函数得到的列别名可用于order by 排序，因为order by执行在select之后。

where, group by, having都不可引用该列，因为这些语句执行在select之前，此时函数尚未计算出值。

--创建测试表
create table te.sc(id int, name varchar(20),class varchar(20), score int);

--给测试表插入数据
insert into te.sc values (1,'张飞','一年一班',100);
insert into te.sc values (2,'刘备','一年一班',99);
insert into te.sc values (3,'李逵','一年一班',95);
insert into te.sc values (4,'小动','一年一班',97);
insert into te.sc values (5,'小智','一年一班',80);
insert into te.sc values (6,'吕布','一年二班',67);
insert into te.sc values (7,'赵云','一年二班',90);
insert into te.sc values (8,'典韦','一年二班',89);
insert into te.sc values (9,'关羽','一年二班',70);
insert into te.sc values (10,'马超','一年二班',98);
insert into te.sc values (11,'张媛','一年一班',100);

--列出每个班分数排名前三的学生
    select * from (select id, name, class, score , 
                row_number() over (partition by class order by score desc) as r1,
                rank() over (partition by class order by score desc) as r2 ,
                dense_rank() over (partition by class order by score desc) as r3 from te.sc) B where r1<=3 ;

id     name     class     score     r1   r2   r3
1      张飞    一年一班    100        1    1    1
11     张媛    一年一班    100        2    1    1
2      刘备    一年一班    99         3    3    2
10     马超    一年二班    98         1    1    1
7      赵云    一年二班    90         2    2    2
8      典韦    一年二班    89         3    3    3

这三个函数的区别主要在分数一致的情况下，row_number()不重复排序，rank()重复且跳数字排序，dense_rank()重复且不跳数字排序。

文章由极客之音整理，本文链接：https://www.bmabk.com/index.php/post/13946.html