Question

有两个表，包表和产品表。就我而言，该软件包包含多个产品。我们需要识别多个产品是否可以匹配已包装记录中的包。下面是一些脚本。

DECLARE @tblPackage TABLE(
    PackageID   int,
    ProductID   int
)

INSERT INTO @tblPackage VALUES(436, 4313)
INSERT INTO @tblPackage VALUES(436, 4305)
INSERT INTO @tblPackage VALUES(436, 4986)

INSERT INTO @tblPackage VALUES(437, 4313)
INSERT INTO @tblPackage VALUES(437, 4305)

INSERT INTO @tblPackage VALUES(442, 4313)
INSERT INTO @tblPackage VALUES(442, 4335)

INSERT INTO @tblPackage VALUES(445, 4305)
INSERT INTO @tblPackage VALUES(445, 4335)

enter image description here

DECLARE @tblProduct TABLE(
    ProductID   int
)

INSERT INTO @tblProduct VALUES(4305) 
INSERT INTO @tblProduct VALUES(4313)

enter image description here

我们有两个产品4305和4313，然后我需要检索匹配的包记录437.只有完全匹配的一个可以返回，所以包436不是正确的。制作多行查询子句并不容易。请有人可以有任何建议吗？感谢。

Answer 1

试试这个。 SQLFIDDLE DEMO

Declare @cnt Int
Select @cnt = count(distinct ProductID) from tblProduct 

SELECT B.packageid 
FROM   (SELECT packageid 
        FROM   tblpackage 
        GROUP  BY packageid 
        HAVING Count(productid) = @cnt) A 
       JOIN tblpackage B 
         ON a.packageid = b.packageid 
WHERE  EXISTS (SELECT 1  FROM tblproduct c WHERE c.productid = b.productid) 
GROUP  BY B.packageid 
HAVING Count(DISTINCT B.productid) = @cnt

Answer 2

这是＆＃34; set-within-sets＆＃34;查询。我会使用聚合和having：

来处理它

select p.PackageID
from @tblPackage p left join
     @tblProduct pr
     on p.ProductId = pr.ProductId
group by p.PackageId
having count(*) = count(pr.ProductId) and
       count(*) = (select count(*) from @tblProduct);

left join保留每个包的所有产品。 having子句中的第一个条件是所有这些产品都与产品表中的产品相匹配。第二个说所有的产品都在那里。

请注意，如果您在任一表格中都有重复项，那么您需要在一个或多个地方使用count(distinct)。您的示例数据表明这不是问题。

Answer 3

这个怎么样：

;WITH MatchingRows
AS
(
    SELECT P1.PackageID, P1.ProductID 
    FROM @tblPackage P1
    INNER JOIN @tblProduct P2
        ON P1.ProductID = P2.ProductID
)
SELECT DISTINCT P1.PackageID
FROM @tblPackage P1
INNER JOIN @tblProduct P2
    ON P1.ProductID = P2.ProductID
WHERE (SELECT COUNT(*) FROM @tblPackage P3 WHERE P3.PackageID = P1.PackageID) /* How may products are in this package */
    = (SELECT COUNT(*) FROM MatchingRows WHERE MatchingRows.PackageID = P1.PackageID) /* how many matching rows in this package - are they the same*/

Answer 4

此查询将有所帮助..

DECLARE @tblPackage TABLE(
    PackageID   int,
    ProductID   int
)

INSERT INTO @tblPackage VALUES(436, 4313)
INSERT INTO @tblPackage VALUES(436, 4305)
INSERT INTO @tblPackage VALUES(436, 4986)

INSERT INTO @tblPackage VALUES(437, 4313)
INSERT INTO @tblPackage VALUES(437, 4305)

INSERT INTO @tblPackage VALUES(442, 4313)
INSERT INTO @tblPackage VALUES(442, 4335)

INSERT INTO @tblPackage VALUES(445, 4305)
INSERT INTO @tblPackage VALUES(445, 4335)

DECLARE @tblProduct TABLE(
    ProductID   int
)

INSERT INTO @tblProduct VALUES(4305) 
INSERT INTO @tblProduct VALUES(4313)

SELECT A.PackageID FROM
    (SELECT PackageID,
           CombinedProductID = 
            (
                SELECT
                    CAST(t2.ProductID AS VARCHAR(100))
                FROM @tblPackage t2
                WHERE
                    t2.PackageID = t1.PackageID
                GROUP BY t2.ProductID
                FOR XML PATH(''), TYPE).value('.', 'varchar(max)'
            )
    FROM @tblPackage t1
    GROUP BY t1.PackageID) AS A
INNER JOIN     
    (SELECT (SELECT ProductID 
         FROM @tblProduct 
         ORDER BY ProductID  
         FOR XML PATH (''), type).value('.', 'varchar(max)') 
     AS CombinedProductID) AS B

ON A.CombinedProductID = B.CombinedProductID

T-SQL如何匹配多行

4 个答案: