Mysql 总和在查询中相乘

我正在做两道作业，几个小时后，我几乎把它们都解决了，最后一个问题是，我的两个查询返回的数值都是双倍的，而不是一个

以下是我所拥有的：

SELECT SUM(P.AMT_PAID) AS TOTAL_PAID, C.CITATION_ID, C.DATE_ISSUED, SUM(V.FINE_CHARGED) AS TOTAL_CHARGED
FROM PAYMENT P, CITATION C, VIOLATION_CITATION V
WHERE V.CITATION_ID = C.CITATION_ID
AND C.CITATION_ID = P.CITATION_ID
GROUP BY C.CITATION_ID;

还有我的另一个：

SELECT C.CITATION_ID, C.DATE_ISSUED, SUM(V.FINE_CHARGED) AS TOTAL_CHARGED, SUM(P.AMT_PAID) AS TOTAL_PAID, SUM(V.FINE_CHARGED) - SUM(P.AMT_PAID) AS TOTAL_OWED
FROM (CITATION C)
LEFT JOIN VIOLATION_CITATION V
ON V.CITATION_ID = C.CITATION_ID
LEFT JOIN PAYMENT P
ON P.CITATION_ID = C.CITATION_ID
GROUP BY C.CITATION_ID
ORDER BY TOTAL_OWED DESC;

我确信我忽略了一些东西。如果有人能告诉我哪里出了差错，那将是一个很大的帮助

Select Sum(P.Amt_Paid) As Total_Paid, C.Citation_Id
    , C.Date_Issued, Sum(V.Fine_Charged) As Total_Charged
From Payment P
    Join Citation C
        On C.Citation_Id = P.Citation_Id
    Join Violation_Citation V
        On V.Citation_Id = C.Citation_Id
Group By C.Citation_Id

首先，应该使用联接语法，而不是使用逗号分隔的表列表。它使它更容易阅读，更标准化，并将有助于防止忽视过滤条款的问题

第二，总和过大的最可能原因是由于连接到

违规_引证

表。如果使用聚合函数删除Group By和列，您可能会看到，

P.AMT\u PAID

对于每个

违规\u引证

实例都会重复。或许，以下方法可以解决问题：

Select Coalesce(PaidByCitation.TotalAmtPaid,0) As Total_Paid
    , C.Citation_Id, C.Date_Issued
    , Coalesce(ViolationByCitation.TotalCharged,0) As Total_Charged
    , Coalesce(ViolationByCitation.TotalCharged,0)
        - Coalesce(PaidByCitation.TotalAmtPaid,0) As Total_Owed
From Citation As C
    Left Join   (
                Select P.Citation_Id, Sum( P.Amt_Paid ) As TotalAmtPaid
                From Payment As P
                Group By P.Citation_Id
                ) As PaidByCitation
        On PaidByCitation.Citation_Id = C.Citation_Id
    Left Join   (
                Select V.Citation_Id, Sum( V.Find_Charged ) As TotalCharged
                From Violation_Citation As V
                Group By V.Citation_Id
                ) As ViolationByCitation
        On ViolationByCitation.Citation_Id = C.Citation_Id

---

Coalesce

的使用是为了确保如果左连接没有为给定的

引文ID

值返回任何行，那么我们将

Null

替换为零。

谢谢你，托马斯，这比我现在的位置稍微高级一些，但实际上它工作得非常好。不过，我会研究它，并将它应用到我的其他一些问题上，因为我一直在讨论这些问题，所有这些问题的最终问题都是价值加倍。