--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);
不管在oracle,还是在8.0版的mysql中,在排序的时候都可以用到三个函数:rank,row_number,dense_rank
--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。