使用 MYSql 计算总和和平均百分比

Question

我有一个Statistics 表，如下所示

parent     key      value   date
MASTER     maths    400     2016-10-12
MASTER     maths    200     2016-10-20
MASTER     maths    350     2016-10-23
MASTER     passed   20%     2016-10-25
MASTER     passed   40%     2016-10-27
MASTER     passed   30%     2016-10-28
MASTER     maths    500     2016-11-12
MASTER     maths    250     2016-11-20 
MASTER     maths    600     2016-11-23 
MASTER     passed   55%     2016-11-25
MASTER     passed   75%     2016-11-27
MASTER     maths    500     2016-12-12
MASTER     maths    250     2016-12-20 
MASTER     maths    600     2016-12-23 
MASTER     passed   70%     2016-12-25
MASTER     passed   65%     2016-12-26
MASTER     passed   45%     2016-12-27
MASTER     maths    500     2017-01-12
MASTER     maths    450     2017-01-20 
MASTER     maths    770     2017-01-23 
MASTER     passed   87%     2017-01-25
MASTER     passed   54%     2017-01-27
MASTER     maths    900     2017-02-12
MASTER     maths    250     2017-02-20 
MASTER     maths    500     2017-02-23 
MASTER     passed   67%     2017-02-25
MASTER     passed   45%     2017-02-27

我需要使用 mysql 显示如下所示的输出

maths  passed   date
950    30%      2016-10
1350   65%      2016-11
1350   60%      2016-12
1720   70.5%    2017-01
1650   56%      2017-02

我写的查询如下

SELECT SUM(CASE WHEN key = 'maths' THEN value ELSE 0 END) AS maths,
       SUM(CASE WHEN key = 'passed' THEN value ELSE 0 END) AS passed 
FROM Statistics
WHERE parent = 'MASTER' GROUP BY MONTH(date) ORDER BY date ASC

但我是通过的总和而不是平均值，也就是说，如果我得到 20%、40% 和 30%，那么我应该得到平均值作为 30%，但对于我上面的查询，我得到了 90%。 % 也消失在结果中

谁能帮我解决这个问题

Answer 1

您必须使用AVG 聚合函数来处理average。

改变：

SUM(CASE WHEN key = 'passed' THEN value ELSE 0 END) AS passed

到：

AVG(CASE WHEN key = 'passed' THEN value ELSE 0 END) AS passed

% 符号被删除

当运算符与不同类型的操作数一起使用时，会发生类型转换以使操作数兼容。一些转换是隐式发生的。例如，MySQL 会根据需要自动将数字转换为字符串，反之亦然。

要显示% 符号，您可以联系它并提供结果。
查询更改后，可能会帮助您获得所需的结果。

mysql> SELECT DATE_FORMAT( `date`, '%Y-%m' ) AS `month`
    ->      , SUM( CASE `key` WHEN 'maths' THEN `value` ELSE NULL END ) AS `maths`
    ->      -- , SUM( CASE `key` WHEN 'passed' THEN `value` ELSE NULL END ) AS `passed`
    ->      , CONCAT( ROUND( AVG( CASE `key` WHEN 'passed' THEN `value` ELSE NULL END ), 2 ), '%' ) AS `passedAvg`
    ->   FROM Statistics
    ->  WHERE parent = 'MASTER'
    ->  GROUP BY MONTH(`date`)
    ->  ORDER BY `date` ASC;

结果可能是这样的：

+---------+-------+-----------+
| month   | maths | passedAvg |
+---------+-------+-----------+
| 2016-10 |   950 | 30.00%    |
| 2016-11 |  1350 | 65.00%    |
| 2016-12 |  1350 | 60.00%    |
| 2017-01 |  1720 | 70.50%    |
| 2017-02 |  1650 | 56.00%    |
+---------+-------+-----------+

文档参考：

• Type Conversion in Expression Evaluation
  • 当运算符与不同类型的操作数一起使用时，会发生类型转换以使操作数兼容。一些转换是隐式发生的。例如，MySQL 会根据需要自动将数字转换为字符串，反之亦然

Answer 2

使用子查询：

要计算 AVG，首先需要知道有多少行“通过”值。

（查看here。）

select yr as `year`, mt as `month`,
       sum(val_maths) as `math`,
       concat(cast((sum(val_passed) / sum(count_passed)) as char(50)), '%')  as `passed`
from (
       select ky, year(dt) as yr, month(dt) as mt,
              if (ky = 'maths', replace(val,'%',''), 0) val_maths,
              if (ky = 'passed', replace(val,'%',''), 0) val_passed,
              if (ky = 'passed', 1, 0) as count_passed    
       from foot
     ) tv
group  by yr, mt
;

| year  | month | maths  | passed |
|-------|-------|--------|--------|
| 2016  | 10    | 950    | 30%    |
| 2016  | 11    | 1350   | 65%    |
| 2016  | 12    | 1350   | 60%    |
| 2017  | 1     | 1720   | 70.5%  |
| 2017  | 2     | 1650   | 56%    |