首先,先建一个测试表
create table praise_record(
id bigint primary key auto_increment,
name varchar(10),
praise_num int
) ENGINE=InnoDB;
复制代码
然后让chatGpt给我们生成几条测试数据
INSERT INTO praise_record (name, praise_num) VALUES ('John', 5);
INSERT INTO praise_record (name, praise_num) VALUES ('Jane', 3);
INSERT INTO praise_record (name, praise_num) VALUES ('Bob', 10);
INSERT INTO praise_record (name, praise_num) VALUES ('Alice', 3);
INSERT INTO praise_record (name, praise_num) VALUES ('David', 7);
INSERT INTO praise_record (name, praise_num) VALUES ('oct', 7);
复制代码
然后就可以开始实现我们的需求:返回点赞的榜单,并返回排名
使用rank()函数返回点赞的榜单, rank() over()
## 注意这里返回的rank字段要用反引号包起来
select name, praise_num, rank() over (order by praise_num desc) as `rank` from praise_record;
+-------+------------+------+
| name | praise_num | rank |
+-------+------------+------+
| Bob | 10 | 1 |
| David | 7 | 2 |
| oct | 7 | 2 |
| John | 5 | 4 |
| Jane | 3 | 5 |
| Alice | 3 | 5 |
+-------+------------+------+
复制代码
可以看到使用rank()函数的时候相同的点赞数会返回相同的排名,排名会产生跳跃,最终的排名不是连续的
使用dense_rank()函数返回点赞的榜单, dense_rank() over()
select name, praise_num, dense_rank() over (order by praise_num desc) as `rank` from praise_record;
+-------+------------+------+
| name | praise_num | rank |
+-------+------------+------+
| Bob | 10 | 1 |
| David | 7 | 2 |
| oct | 7 | 2 |
| John | 5 | 3 |
| Jane | 3 | 4 |
| Alice | 3 | 4 |
+-------+------------+------+
复制代码
与rank()函数相同的是,相同点赞数会返回相同的排名,但是dense_rank()返回的最终排名是连续的排名
row_number()函数返回点赞的榜单,row_number() over()
select name, praise_num, row_number() over (order by praise_num desc) as `rank` from praise_record;
+-------+------------+------+
| name | praise_num | rank |
+-------+------------+------+
| Bob | 10 | 1 |
| David | 7 | 2 |
| oct | 7 | 3 |
| John | 5 | 4 |
| Jane | 3 | 5 |
| Alice | 3 | 6 |
+-------+------------+------+
复制代码
row_number()函数适合当返回的列表只需要序号时使用
以上三个函数都是MySQL8.0新加入的,所以在MySQL5.7这些老版本上我们可以模拟实现一下,顺便学习一下这三个窗口函数的实现原理
select p1.name, p1.praise_num, count(p2.praise_num) + 1 as `rank` from praise_record p1
left join praise_record p2 on p1.praise_num < p2.praise_num
group by p1.name, p1.praise_num
order by `rank`;
+-------+------------+------+
| name | praise_num | rank |
+-------+------------+------+
| Bob | 10 | 1 |
| David | 7 | 2 |
| oct | 7 | 2 |
| John | 5 | 4 |
| Jane | 3 | 5 |
| Alice | 3 | 5 |
+-------+------------+------+
复制代码
我们可以使用自联接的方式将每个分数低于当前行分数的记录计数,最后将计数值加1作为当前行的排名,来模拟实现rank()
select p1.name, p1.praise_num, count(distinct p2.praise_num) + 1 as `dense_rank` from praise_record p1
left join praise_record p2 on p1.praise_num < p2.praise_num
group by p1.name, p1.praise_num
order by `dense_rank`;
+-------+------------+------------+
| name | praise_num | dense_rank |
+-------+------------+------------+
| Bob | 10 | 1 |
| oct | 7 | 2 |
| David | 7 | 2 |
| John | 5 | 3 |
| Jane | 3 | 4 |
| Alice | 3 | 4 |
+-------+------------+------------+
复制代码
dense_rank的实现与rank差不多,唯一的区别是增加了distinct对点赞数做了去重,这样子对不同的点赞数返回的排名就是连续的
##使用自定义变量得先初始化
set @rowNum = 0;
select name, praise_num, @rowNum := @rowNum +1 as `row_number` from praise_record order by praise_num desc ;
+-------+------------+------------+
| name | praise_num | row_number |
+-------+------------+------------+
| Bob | 10 | 1 |
| David | 7 | 2 |
| oct | 7 | 3 |
| John | 5 | 4 |
| Jane | 3 | 5 |
| Alice | 3 | 6 |
+-------+------------+------------+
复制代码
我们可以使用一个rowNum变量来记录行号,每一行的数据rowNUm都+1,这样子就可以得到我们想要的序号
1.rank()函数返回的排名会产生跳跃
2.dense_rank()函数返回的排名是连续的
3.row_number()函数返回的排名类似序号
4.窗口函数是MySQL8.0新增的特性,如果在低版本的MySQL要自己模拟实现一下
本文系转载,前往查看
如有侵权,请联系 cloudcommunity@tencent.com 删除。
本文系转载,前往查看
如有侵权,请联系 cloudcommunity@tencent.com 删除。