我正在使用 MYSQL 计算学生表中每个科目的 TOP-N。

Question

我有一个学生表，其中包含 student_id、分数和科目

CREATE TABLE IF NOT EXISTS students 
(student_id INT(3),   subject ,VARCHAR(45),   score INT(3) );

插入的数据是

insert into students values(1,'math',70);
insert into students values(1,'science',71);
insert into students values(1,'history',72);
insert into students values(1,'english',73);
insert into students values(1,'kannada',74);
insert into students values(3,'math',50);
insert into students values(3,'science',51);
insert into students values(3,'history',52);
insert into students values(3,'english',53);
insert into students values(3,'kannada',54);
insert into students values(2,'math',60);
insert into students values(2,'science',61);
insert into students values(2,'history',62);
insert into students values(2,'english',63);
insert into students values(2,'kannada',64);

使用查询后我得到了所需的输出，

select student_id,score,subject
    from
        (select @prev := '', @n:=0) init
    join
        (select @n := if(subject != @prev , 1, @n+1) as n,
                 @prev := subject,
                 student_id,score,subject from students
                 order by
                 subject asc, 
                 score desc
        ) x
        where n<=2
        order by subject, score desc;

我只是不明白这是如何工作的，为什么需要加入？这是子查询吗？ from 子句中的语句会在每一行数据上运行吗？有人请给我解释一下。我正在学习 SQL。

注意：我在网上找到了与此类似的查询，我只是根据我的要求对其进行了调整，这不是我的工作。

Answer 1

只需要连接，以便您可以在查询中初始化变量@prev 和@n。这需要与您尝试过滤的查询分开完成。您可以改为在查询之前执行此操作，但这会将所有内容放在一个独立的查询中。

SET @prev = '';
SET @n = 0;
SELECT student_id, score, subject
FROM 
    (select @n := if(subject != @prev , 1, @n+1) as n,
             @prev := subject,
             student_id,score,subject from students
             order by
             subject asc, 
             score desc
    ) x
where n<=2
order by subject, score desc;

在这种情况下，使用子查询，以便您可以选择所需的 n <= 2 行。