CASE语句仍然返回多行

Question

以下查询每个ID返回两行.

select distinct ID,
count (COURSE_IDENTIFICATION) as tot_course_id,

COUNT(case when course_level = 'G'
     then (COURSE_IDENTIFICATION)
     else '0'end) count_grad,

COUNT (case when course_level = 'U'
       then (COURSE_IDENTIFICATION)
       else '0'end) count_under

from ods.STUDENT_COURSE

where ACADEMIC_PERIOD = '201890'
and COURSE_BILLING_CREDITS >0.0
group by ID, course_level
order by 1

我收到以下结果:

ID      |TOT_COURSE_NUM |COUNT_GRAD |COUNT_UNDER
-------------------------------------------------
12345   |1              |1          |1
12345   |2              |2          |2

有没有办法确保同一行中的grad和undergrad总数都计算在内？

Answer 1

您的直接问题是,course_level在group by.第二个问题count()是不正确.我认为这应该足够了:

select ID,
       count(*) as tot_course_id,
       sum(case when course_level = 'G' then 1 else 0 end) as count_grad,
       sum(case when course_level = 'U' then 1 else 0 end) as count_under
from ods.STUDENT_COURSE sc
where ACADEMIC_PERIOD = '201890' and
      COURSE_BILLING_CREDITS > 0.0
group by ID
order by 1;

count(<expression>)计算非NULL值的数量.你else '0'仍然是非空的,所以它被计算在内. sum()是一种处理这个问题的简单方法.

这个版本假设COURSE_IDENTIFICATION不是NULL.这似乎是一个名为的合理假设STUDENT_COURSE.它很容易修改以处理NULL值.

并且,select distinct几乎从不适合group by.