Question

所以这里有一些样本信息：

ROW  |     PHONE     |   ID   |  NAME
_________________________________________
1      000-000-0000      3     Company1
2      000-000-0000     38     Company2
1      123-456-7890      0     Company3
2      123-456-7890      0     Company4
3      123-456-7890      0     Company5
1      197-865-4329      0     Company6
2      197-865-4329     18     Company7
1      654-987-1230     74     Company8
2      654-987-1230      2     Company9
1      951-753-8462      2     Company10
2      951-753-8462      3     Company11
3      951-753-8462      3     Company12
4      951-753-8462      3     Company13

使用以下代码从表中提取此信息：

USE MyTable
GO

SELECT ROW_NUMBER() OVER (PARTITION BY AI.telephone1 ORDER BY AI.telephone1) AS Row
        ,AI.telephone1 AS [Main Phone #]
        ,AI.new_id AS [ID]
        ,AI.name AS [Account Name]
        ,AI.emailaddress1 AS [Email Address]
        ,AI.contactname AS [Primary Contact]
FROM AccountsInfo AI
WHERE AI.telephone1 IN (
        SELECT telephone1
        FROM MyDatabase.dbo.AccountsInfo AI
        WHERE telephone1 != 'NULL'
            AND telephone1 != '--'
        GROUP BY telephone1
        HAVING COUNT(*) > 1
        )
ORDER BY telephone1

我现在需要做的是，但无法绕过我的脑袋，遍历每个分区并仅提取具有匹配ID号的结果。请参阅下文，了解我希望桌子最终看起来像什么......

    PHONE    |   ID   |   NAME
_________________________________________
123-456-7890      0      Company3
123-456-7890      0      Company4
123-456-7890      0      Company5
951-753-8462      3      Company11
951-753-8462      3      Company12
951-753-8462      3      Company13

正如您所看到的，公司3,4和5进入了结果，因为他们三个都有匹配的ID号和匹配的电话号码。公司11,12和13制作了它们的分区，因为它们具有匹配的ID号。公司10没有被拉，因为它的ID号不同于11,12和13。

我需要做的是让代码迭代每个分区（按电话号码分区）并检查分区中每行的ID号与分区中的其他行。最终结果将只是在其分区内的其他位置重复ID号的行。

如果您需要任何其他信息，请与我们联系。

Answer 1

这个怎么样？

SELECT   AI.telephone1 AS [Main Phone #]
        ,AI.new_id AS [ID]
        ,AI.name AS [Account Name]
        ,AI.emailaddress1 AS [Email Address]
        ,AI.contactname AS [Primary Contact]
FROM AccountsInfo AI
INNER JOIN 
(
        SELECT telephone1,new_id
        FROM MyDatabase.dbo.AccountsInfo AI
        WHERE telephone1 != 'NULL'
            AND telephone1 != '--'
        GROUP BY telephone1, new_id
        HAVING COUNT(*) > 1
) T
ON AI.telephone1 = T.telephone1
AND AI.new_id = T.new_id
ORDER BY AI.telephone1

Answer 2

您也可以使用having：

select
    "phone", "id"
from
    "data"
group by "phone", "id"
having ( min ( "row" ) < max( "row" ));

有关更多信息和可能性，请查看Joe Celko's Thinking in Sets

请参阅SQL-Fiddle

SQL - 尝试迭代分区以查找重复项

2 个答案: