具有连接到三个表的SQL子查询

Question

我需要查询3张桌子。以下是示例：

学生表格：

 StudentId     Name
 --------------------
    1           Aaa
    2           Bbb
    3           Ccc

ClassStudent表格

 ClassId       StudentId
 -----------------------
    1             1
    1             2
    2             3

付款表格：

PaymentTime   ClassId
---------------------
08.10.2019       1
08.10.2019       2
08.09.2019       1
08.09.2019       2
08.08.2019       1
08.07.2019       1
08.06.2019       1

我需要这样：从Payment表中，我想要PaymentTime大于08.09.2019的人数。

我该如何做

SELECT ClassStudent.ClassId, COUNT(ClassStudnet.ClassId) AS StudentCount
FROM ClassStudent CS
INNER JOIN PaymentTable PT ON PT.ClassId = CS.ClassId 
AND WHERE PT.PaymentTime > CONVERT(date,'08.09.2019')
GROUP BY CS.ClassId

如何在Class表中查找对Payment的查询数量？

我希望分别在“ 08.09.2019”之前和之后付款的人数。

赞：

ClassId TotalStudent PaymentBefore08.09.2019 PaymentAfter08.09.2019
--------------------------------------------------------------------
 1            10                3                       7  
 2            15                4                      11

Answer 1

    SELECT 
COUNT(CASE WHEN PT.PaymentTime > CONVERT(date, '08.09.2019') THEN COUNT(PT.PaymentTime) END) AS 'AfterTotalCount08.09.2019',
COUNT(CASE WHEN PT.PaymentTime < CONVERT(date, '08.09.2019') THEN COUNT(PT.PaymentTime) END) AS 'BeforeTotalCount08.09.2019'
    FROM PaymentTable PT

我想要这样

Answer 2

[如果我的解释正确，ClassStudent和Student表似乎都是无关紧要的，因为您似乎是在计算付款而不是学生。您也许可以这样做：

unested

SELECT PT.ClassId, COUNT(*) AS PaymentCount
FROM PaymentTable PT
WHERE PT.PaymentTime > CONVERT(date, '08.09.2019')
GROUP BY PT.ClassId

但是这不算学生。它计算付款次数（因此为PaymentCount而不是StudentCount）。如果一位学生可以为1堂课支付2笔款项，那么在上面的查询中这将算作2笔。要实际计算学生人数，您需要在付款与学生之间建立直接关系，该关系在您提供的表格中不存在。
根据您的评论修改的答案：
SELECT T1.ClassID, T1.TotalStudents,
   isnull(T2.PaymentCountAfter, 0) PaymentCountAfter,
   isnull(T2.PaymentCountBeforeEqual, 0) PaymentCountBeforeEqual
FROM (
   SELECT CS.ClassId, COUNT(*) TotalStudents
   FROM ClassStudent CS
   GROUP BY CS.ClassId
) T1
LEFT OUTER JOIN (
   SELECT PT.ClassId,
      COUNT(CASE WHEN PT.PaymentTime > CONVERT(date, '08.09.2019') THEN 1) AS PaymentCountAfter,
      COUNT(CASE WHEN PT.PaymentTime <= CONVERT(date, '08.09.2019') THEN 1) AS PaymentCountBeforeEqual
   FROM PaymentTable PT
   GROUP BY PT.ClassId
) T2 on T2.ClassId  = T1.ClassId

这可能是通过Windows函数而不是左联接子查询来完成的。