具有各自值的多列到行

Question

试图将多列放入行中。 预期的结果是

这是我尝试过的示例数据。 如果整体速度更快，我也愿意接受 unpivot。 完整数据有 15 个 AttributeID、AttributeData 列。

DROP TABLE Attribute;

CREATE TABLE Attribute 
(
        Producttitle varchar(200),  
        AttributeID_1 varchar(50),  
        AttributeData_1 varchar(50), 
        AttributeID_2   varchar(50),
        AttributeData_2 varchar(50),
        AttributeID_3 varchar(50),  
        AttributeData_3 varchar(50)
);

INSERT INTO Attribute 
VALUES ('title1', '3145', 'Specific', '30', 'Yes', '40', 'Pink')

INSERT INTO Attribute 
VALUES ('title2', '17', 'Stainless', '19', 'smoke', '19', 'Something');

SELECT
    Producttitle, 
    [AttributeID],
    [AttributeData] 
FROM
    Attribute
CROSS APPLY
    (SELECT 'Indicator1', [AttributeID_1] UNION ALL
     SELECT 'Indicator2', [AttributeID_2] UNION ALL
     SELECT 'Indicator3', [AttributeID_3]) c (indicatorname, [AttributeID])
CROSS APPLY
    (SELECT 'Indicator1', [AttributeData_1] UNION ALL
     SELECT 'Indicator2', [AttributeData_2] UNION ALL
     SELECT 'Indicator3', [AttributeData_3]) d (indicatorname, [AttributeData]);

Answer 1

您可以使用cross apply来取消透视数据集。 values()更简单：

select a.title, x.*
from attribute a
cross apply (values
    (a.attributeId_1, a.attributeData_1),
    (a.attributeId_2, a.attributeData_2),
    (a.attributeId_3, a.attributeData_3)
) as x(attributeId, attributeData)

请注意，这是有效的，因为两组列具有一致的数据类型 - 否则将需要额外的转换。

Answer 2

GMB 的解决方案真的很酷，但基本工会也可以：

SELECT Producttitle, AttributeID_1 AttributeID, AttributeData_1 AttributeData
from attribute
union
SELECT Producttitle, AttributeID_2 AttributeID, AttributeData_2 AttributeData
from attribute
union
SELECT Producttitle, AttributeID_3 AttributeID, AttributeData_3 AttributeData
from attribute