Question

我有问题。我有4张桌子：

Invoice_Payment，Invoice，Client和Calendar

基本上，我有以下查询，它运行良好，除了没有date_due的月份不返回。 I.E只会返回date_due的月份。

注意：日历表只列出一年中的每一天。它包含一个col调用date_field

数据库：http://oberto.co.nz/jInvoice.txt

预期输出：我当前的查询将返回如下内容：

month       total
August      5
September   0
October  196
November  205
December  214
January  229

请注意九月没有退回？这是因为表Invoice_Payment没有date_due记录

我想我必须使用左连接并加入日历表，例如： LEFT JOIN日历在Invoice_Payments.date_paid = Calendar.date_field。但我没有运气

SELECT MONTHNAME(Invoice_Payments.date_paid) as month, SUM(Invoice_Payments.paid_amount) AS total
FROM Invoice, Client, Invoice_Payments
WHERE Client.registered_id = 1
AND Client.id = Invoice.client_id
And Invoice.id = Invoice_Payments.invoice_id
AND date_paid IS NOT NULL
GROUP BY YEAR(Invoice_Payments.date_paid), MONTH(Invoice_Payments.date_paid)

任何帮助表示感谢。

Answer 1

听起来您正在尝试查找范围内所有日期的值，无论是否存在值。假设我们有一个类似的Calendar表：

Create Table Calendar
(
    [Date] not null Primary Key
)

您的查询可能如此（其中X和Y代表您正在调查的范围的开始和结束日期）：

Select Year(C.Date), MonthName(C.Date) As Month
    , Coalesce(Sum(IP.paid_amount),0) As Total
From Calendar As C
    Left Join (Invoice As I
        Join Client As C1
            On C1.id = I.client_id
                And C.registered_id = 1
        Join Invoice_Payments As IP
            On IP.Invoice_id = I.Id)
        On IP.date_paid = C.date
Where C.Date Between X and Y
Group By Year(C.Date), MonthName(C.Date)

从技术上讲，上面的查询应该可以解决问题。但是，另一种方法是使用您在评论中查询的派生表：

Select Year(C.Date), MonthName(C.Date) As Month
    , Coalesce(Sum(Z.paid_amount),0) As Total
From Calendar As C
    Left Join   (
                Select IP.date_paid, IP.paid_amount
                From Invoice As I
                        Join Client As C1
                            On C1.id = I.client_id
                                And C.registered_id = 1
                        Join Invoice_Payments As IP
                            On IP.Invoice_id = I.Id
                ) As Z
        On Z.date_paid = C.date
Where C.Date Between X and Y
Group By Year(C.Date), MonthName(C.Date)

Answer 2

我遇到了类似的问题并通过将非日历表条件移出WHERE子句并进入连接来解决它。请在此处查看我的完整说明：https://gist.github.com/1137089

在你的情况下，你需要将from table1，table2，table3分解为x = y上的左连接table1 ......

Answer 3

为了便于阅读，我会明确定义所有联接。

外连接的一个简单示例，它保留第一个表中的所有行，用空值填充第二个表中任何缺少的列：

select column1,
from table1
left outer join table2
  on table1.key = table2.key

假设Calendar表是导致您丢失数据的表，这里是您的查询的一个镜头：

SELECT MONTHNAME(Invoice_Payments.date_paid) as month,
       SUM(Invoice_Payments.paid_amount)
FROM Invoice
JOIN Client
  ON Client.id = Invoice.client_id
 AND Client.registered_id = 1
JOIN Invoice_Payments
  ON Invoice_Payments.invoice_id = Invoice.id
LEFT OUTER JOIN Calendar
  ON Calendar.key = #yourothertable#.key    
WHERE date_paid IS NOT NULL
GROUP BY YEAR(Invoice_Payments.date_paid), MONTH(Invoice_Payments.date_paid)

祝你好运！

mySQL Left加入有条件的多个表

3 个答案: