Question

我正在查看上一篇关于＆＃34;部门使用count＆＃34;似乎在分开的问题上有问题。下面是我的查询，我得到了结果，但结果是错误的，因为＆＃34;除以＆＃34; count(*)似乎使用了所有商店的数量，而不是商店分组。

我感兴趣的是＆＃34;除以＆＃34;还应该基于该特定商店的计数数量。我对分频的查询的第二部分应该是＆＃34; select count(*) *100 group by stores＆＃34;

SELECT store, CAST(((  
    SUM(CASE WHEN SCORE BETWEEN 9 AND 10 THEN 1 ELSE 0 END) * 1.0 -
    SUM(CASE WHEN SCORE BETWEEN 0 AND 6 THEN 1 ELSE 0 END)) / COUNT(*) * 100) AS signed) AS NPS
FROM npsnetsuite
WHERE SCORE IS NOT NULL group by STORE;

示例：

Store A has got total count of 7 rows
Store B has got total count of 12 rows

该部门应该是

XX/7 for Store A
XX/12 for Store B

示例数据：

Date       Store    Score   Case When between 9-10  case when 0-6
18/04/2016  A         9                  1  
20/04/2016  A         8     
26/04/2016  A         10                 1  
29/04/2016  A         8     
30/04/2016  A        10                  1  
01/05/2016  A        10                  1  
06/05/2016  A        8      
11/05/2016  A        10                  1  
23/05/2016  A        10                  1  
04/06/2016  A         6                               1
09/06/2016  A         8     
08/06/2016  A        10                  1  
05/04/2016  B         4                               1 
07/04/2016  B        10                  1  
18/04/2016  B        10                  1  
27/04/2016  B         8     
01/05/2016  B         9                  1  
13/05/2016  B        10                  1  
13/06/2016  B         1                              1
14/06/2016  B         6                              1

示例数据：

期望输出：

Store A 50
Store B 13

基本上Store A is (7-1)/12 * 100 = 50。

7是分数为9 to 10的情况下的计数，1是分数0 to 6的情况下的计数。所以我得到它的第一部分是每个商店的得分数为9到10分数0到6的数量。我需要除以特定商店的计数。如上例所示，商店A包含12行，商店B包含8行。我当前的查询是计算所有商店的数量。

谢谢

Answer 1

谢谢大家〜最后，设法通过此查询获得上述结果:)：

select a.STORE, round((a.count_one / b.count_two) * 100) as final_count from 
(select SCORE, STORE, SUM(CASE WHEN SCORE BETWEEN 9 AND 10 THEN 1 ELSE 0 END) * 1.0 -
  SUM(CASE WHEN SCORE BETWEEN 0 AND 6 THEN 1 ELSE 0 END) as count_one from npsnetsuite where SCORE IS NOT NULL group by STORE) a,
(select SCORE, STORE, count(*) as count_two from npsnetsuite where  SCORE IS NOT NULL group by STORE) b
where a.STORE = b.STORE and a.SCORE IS NOT NULL;

使用MySQL进行分组

1 个答案: