在SQL Server 2012中使用CTE进行多行连接

输入：

所需输出： 连接每个要素行组并更新产品标签 条件：如果产品标签中不存在最后一个，则添加功能标签

if object_id('tempdb..#Product') is not null drop table #Product
if object_id('tempdb..#ProductFeatures') is not null drop table #ProductFeatures
create table #Product(Product_id int, Product_label nvarchar(200))
create table #ProductFeatures(Product_id int, FeatureId int, Feature_label nvarchar(30), Feature_Value nvarchar(30))

insert into #Product(Product_id, Product_label) values
 (1, 'HL Mountain Frame - Silver, 42')
,(2, 'Road-650 Black, 58')

insert into #ProductFeatures(Product_id, FeatureId, Feature_label, Feature_Value) values
 (1, 1, 'Color', 'Silver')
,(1, 2, 'Height', '42')
,(1, 3, 'Widht',  '12')

,(2, 1, 'Color', 'Black')
,(2, 2, 'Height', '58')
,(2, 3, 'Widht', '58')

例如，不会添加“银”，因为它存在于产品标签中

if object_id('tempdb..#Product') is not null drop table #Product
if object_id('tempdb..#ProductFeatures') is not null drop table #ProductFeatures
create table #Product(Product_id int, Product_label nvarchar(200))
create table #ProductFeatures(Product_id int, FeatureId int, Feature_label nvarchar(30), Feature_Value nvarchar(30))

insert into #Product(Product_id, Product_label) values
 (1, 'HL Mountain Frame - Silver, 42')
,(2, 'Road-650 Black, 58')

insert into #ProductFeatures(Product_id, FeatureId, Feature_label, Feature_Value) values
 (1, 1, 'Color', 'Silver')
,(1, 2, 'Height', '42')
,(1, 3, 'Widht',  '12')

,(2, 1, 'Color', 'Black')
,(2, 2, 'Height', '58')
,(2, 3, 'Widht', '58')

问题:

---

如何计算CTE（而不是使用光标）？

例如，您可以使用XML的

来实现这一点

Product_id  Product_label
----------- ------------------------------------
1           HL Mountain Frame - Silver, 42 12
2           Road-650 Black, 58

达卡的解决方案非常好。我想提供另一种方法，使用pivot：

UPDATE P
 SET P.Product_label = P2.newValue
FROM #Product AS P
INNER JOIN (
SELECT IP.Product_ID, IP.Product_label +
ISNULL(

        (SELECT ' ' + PF.Feature_value AS [text()]
        FROM #ProductFeatures AS PF
        WHERE PF.Product_id = IP.Product_id
        AND IP.Product_label NOT LIKE '%' + PF.Feature_value + '%'
        FOR XML PATH('')
    ), '') AS newValue
FROM #Product AS IP
) AS P2
ON P.Product_ID = P2.Product_ID

注：以上解决方案未优化。。。但我提到这是另一种方式

干杯，

Maciej
到@Darka，
谢谢你的帮助

只需在脚本中进行一些调整，因为我们要搜索单词：）

Sql Server游标，但我认为还有更简单的解决方案是的，没错，但OP想更新产品标签；）使用pivot会很好，但每个产品的功能集并不相同，我们不能为DynamicQueryPivot列出它们。
charindex(' ' + PF.Feature_value, IP.Product_label) = 0

SELECT IP.Product_ID, IP.Product_label +
ISNULL(

        (SELECT ' ' + PF.Feature_value AS [text()]
        FROM #ProductFeatures AS PF
        WHERE PF.Product_id = IP.Product_id
        AND charindex(' ' + PF.Feature_value, IP.Product_label) = 0
        FOR XML PATH('')
    ), '') AS newValue
FROM #Product AS IP