从SQL中选择多行到子查询中的一行

Question

我有这样的查询:

DECLARE @razem VARCHAR(MAX);

SELECT  Ordering.orderID , 
        Document.number, 
        (User_info.name  +' '+ User_info.surname),   
        Ordering.dateStart, 
        Ordering.dateEnd ,   
        (
            select   COALESCE(' ',@razem)+sell_type.name as r  
            from    Ordering_sell_type, Sell_type 
            where   orderID = Ordering.orderID and 
                    Ordering_sell_type.sell_typeID = sell_type.sell_typeID
        ) podz
FROM    Ordering, User_info, Product_Document, Document, Document_type   
WHERE   Ordering.orderID = Product_document.orderID  
        AND Document.documentID = Document_type.documentID  
        AND Document.documentID = Product_document.documentID  
        AND  Ordering.userID = User_info.userID   
        AND Ordering.isClosed = 1 AND Document_type.typeID = 1   
GROUP   BY  Document.isitfiscal, Document.refDocID, 
            Document.number, Ordering.orderID, User_info.name, 
            User_info.surname, Ordering.dateStart, 
            Ordering.dateEnd , Ordering.isCLosed  
ORDER   BY Ordering.dateEnd

在COALESCE函数中,我想获得所选订单的所有付款类型 - 例如,orderID 123有payTypes =卡,现金,orderID有payTypes =现金.

问题是,我想有它在一个简单的行作为主查询,像最后一行:订单,Document.number,UserInfo.name +姓,dateStart,dateEnd, - >卡,现金< - 而是试图像上面的查询后,我收到了错误:

Subquery returned more than 1 value. This is not permitted when the subquery follows =, !=, <, <= , >, >= or when the subquery is used as an expression.

因为它返回多行.是否有可能在子查询中获取付款类型并返回为一个字符串？

Answer 1

根据您使用的语法,我假设您使用的是SQL-Server,因此您可以使用SQL-Servers XML扩展来连接字符串.

SELECT  Ordering.orderID, 
        Document.number, 
        [UserName] = User_info.name  +' '+ User_info.surname,   
        Ordering.dateStart, 
        Ordering.dateEnd,
        [podz] = STUFF((    SELECT  DISTINCT ' ' + SellType.Name
                            FROM    Ordering_Sell_Type
                                    INNER JOIN Sell_Type
                                        ON Sell_Type.sell_typeID = Ordering_Sell_Type.sell_typeID
                            WHERE   Ordering.OrderID = Ordering_SellType.OrderId
                            FOR XML PATH(''), TYPE
                        ).value('.', 'NVARCHAR(MAX)'), 1, 1, '')
FROM    Ordering
        INNER JOIN User_Info
            ON Ordering.UserID = User_Info.UserID
        INNER JOIN ProductDocument
            ON Ordering.OrderID = Product_Document.OrderID
        INNER JOIN Document
            ON Document.DocumentID = Product_Document.DocumentID
        INNER JOIN Document_Type
            ON Document_Type.DocumentID = Document.DocumentID
WHERE   Ordering.IsClosed = 1
AND     Document_Type.TypeID = 1
ORDER BY Ordering.dateEnd;

注意我已经用ANSI 92替换了所有ANSI 89连接,因为这是更现代的语法,并且通常被认为是更易读的选项(我说通常被接受,因为它当然是个人偏好,并且还有一些情况Oracle优化ANSI89连接优化).

编辑

看过您的数据重复项来自Product_Document表,您可以使用以下方法删除它们:

SELECT  Ordering.orderID, 
        Document.number, 
        [UserName] = User_info.name  +' '+ User_info.surname,   
        Ordering.dateStart, 
        Ordering.dateEnd,
        [podz] = STUFF((    SELECT  DISTINCT ' ' + SellType.Name
                            FROM    Ordering_Sell_Type
                                    INNER JOIN Sell_Type
                                        ON Sell_Type.sell_typeID = Ordering_Sell_Type.sell_typeID
                            WHERE   Ordering.OrderID = Ordering_SellType.OrderId
                            FOR XML PATH(''), TYPE
                        ).value('.', 'NVARCHAR(MAX)'), 1, 1, '')
FROM    Ordering
        INNER JOIN User_Info
            ON Ordering.UserID = User_Info.UserID
        INNER JOIN 
        (   SELECT  DISTINCT OrderID, DocumentID
            FROM    Product_Document
        ) Product_Document
            ON Ordering.OrderID = Product_Document.OrderID
        INNER JOIN Document
            ON Document.DocumentID = Product_Document.DocumentID
        INNER JOIN Document_Type
            ON Document_Type.DocumentID = Document.DocumentID
WHERE   Ordering.IsClosed = 1
AND     Document_Type.TypeID = 1
ORDER BY Ordering.dateEnd;