如何计算 MySQL 中每个组的 运行 总和

How to calculate running sum for each group in MySQL

我可以使用 sum() 函数进行直接求和。但我这里有不同的情况。我有一个只有 2 个字段的 table Student。例如假设整个 class:

只有 1 个学生
CREATE TABLE student
    (`dateOfExam` date, score int)
;
    
INSERT INTO student
    (`dateOfExam`, `score`)
VALUES
    ('2020-05-28',5),
    ('2020-05-29',5),
    ('2020-05-30',10),
    ('2020-06-03',10),
    ('2020-06-05',5),
    ('2020-07-21',20),
    ('2020-07-22',10),
    ('2020-07-28',10)
;

我有他的 分数 参加考试的日子,在运行时间中多了一列,即举行考试的月份:

查询是(昨天从 Whosebug 获得帮助):

select date_format(dateOfExam, '%Y-%m') ExamMonth
     , dateOfExam
     , score 
  from student;

结果:

+-----------+------------+-------+
| ExamMonth | dateOfExam | score |
+-----------+------------+-------+
| 2020-05   | 2020-05-28 |     5 |
| 2020-05   | 2020-05-29 |     5 |
| 2020-05   | 2020-05-30 |    10 |
| 2020-06   | 2020-06-03 |    10 |
| 2020-06   | 2020-06-05 |     5 |
| 2020-07   | 2020-07-21 |    20 |
| 2020-07   | 2020-07-22 |    10 |
| 2020-07   | 2020-07-28 |    10 |
+-----------+------------+-------+

我的要求是每个月都要奖励这个学生。我会继续为每个月的每个日期添加他的分数,当累计分数达到 10 时给他 Reward1,当累计分数达到 Reward2总和达到 20。所以最后的 table 应该是这样的:

+---------------+---------------+-------+---------------+---------------+
| ExamMonth     |  dateOfExam   | Score |    Reward1    |   Reward2     |
+---------------+---------------+-------+---------------+---------------+
|    2020-05    |  2020-05-28   |   5   |               |               |
|               |  2020-05-29   |   5   |       Y       |               |
|               |  2020-05-30   |   10  |               |       Y       |
|---------------|---------------|-------|---------------|---------------|
|    2020-06    |  2020-06-03   |   10  |       Y       |               |
|               |  2020-06-05   |   5   |               |               |
|---------------|---------------|-------|---------------|---------------|
|    2020-7     |  2020-07-21   |   20  |       Y       |       Y       |
|               |  2020-07-22   |   10  |               |               |
|               |  2020-07-28   |   10  |               |               |
+---------------+---------------+-------+---------------+---------------+

奖励字段可以是布尔值,空奖励行可以设置为 N 或 False 或任何看起来符合逻辑的值。 这没有帮助: Calculate running sum

请帮我实现这个 objective。建议一些方法。

这里是fiddle.

以下代码片段按 ExamMonth 上的基本查询分组,然后在决定 Reward1 和 Reward2 的值时使用一个案例。此查询仅供参考。请重写为最适合你的。


select DERIVED2.ExamMonth, CASE WHEN DERIVED2.Cumul_Score >= 10 THEN 'Y'

ELSE ''
END AS Rewards1,
CASE WHEN DERIVED2.Cumul_Score >= 20 THEN 'Y'

ELSE ''
END AS Rewards2
FROM 
(
select DERIVED1.ExamMonth, SUM(DERIVED1.score) as Cumul_Score 
FROM
(
select date_format(dateOfExam, '%Y-%m') ExamMonth,
       dateOfExam, score
from student
order by dateOfExam
) DERIVED1

GROUP BY ExamMonth

) DERIVED2

首先计算 CTE 中每个月的 运行 分数总和。
然后应用您的条件:

with cte as (
  select date_format(dateOfExam, '%Y-%m') ExamMonth,
         dateOfExam, score, 
         sum(score) over (partition by date_format(dateOfExam, '%Y-%m') order by dateOfExam) total
  from student
)
select ExamMonth, dateOfExam, score, 
       case when sum(total >= 10) over (partition by ExamMonth order by dateOfExam) = 1 then 'Y' end Reward1,
       case when sum(total >= 20) over (partition by ExamMonth order by dateOfExam) = 1 then 'Y' end Reward2
from cte

参见demo
结果:

> ExamMonth | dateOfExam | score | Reward1 | Reward2
> :-------- | :--------- | ----: | :------ | :------
> 2020-05   | 2020-05-28 |     5 | null    | null   
> 2020-05   | 2020-05-29 |     5 | Y       | null   
> 2020-05   | 2020-05-30 |    10 | null    | Y      
> 2020-06   | 2020-06-03 |    10 | Y       | null   
> 2020-06   | 2020-06-05 |     5 | null    | null   
> 2020-07   | 2020-07-21 |    20 | Y       | Y      
> 2020-07   | 2020-07-22 |    10 | null    | null   
> 2020-07   | 2020-07-28 |    10 | null    | null