所以我有一个有趣的小循环，寻找重复的地址。我终于调整运行相当快，但我不能完全得到我需要从它。
样本数据和预期结果
循环如下：

WHILE @cnt <= @max
    BEGIN
        SELECT
            @customer_id=[customer_id],
            @add1   =[add1], 
            @add2   =[add2], 
            @add3   =[add3], 
            @zip    =[zip]
          FROM #tmpCustomers
          WHERE [id] = @cnt

        -- Ensures customer_id is not already in the dupeCustomerAddress table
        IF NOT EXISTS( SELECT [customer_id] FROM dupeCustomerAddress WHERE [customer_id] = @customer_id )
        BEGIN

            INSERT INTO dupeCustomerAddress
                SELECT
                    [customer_id],
                    [address1],
                    [add1],
                    [address2],
                    [add2],
                    [address3],
                    [add3],
                    [zip]
                  FROM [working_customerAddress]
                  WHERE
                    -- Removes the Record used for comparison
--                  [customer_id] != @customer_id AND
                    -- Don't need to include records already processed
                    [customer_id] > @cnt AND
                    (
                        (       -- Address Line 1 or 2 matches Comparison Line 1 AND Zip matches
                            (
                              ( [address1] IS NOT NULL AND [add1] = @add1 ) OR  
                              ( [address2] IS NOT NULL AND [add2] = @add1 ) 
                            ) AND [zip] = @zip
                        )
                        OR
                        (       -- Address Line 1 or 2 matches Comparison Line 2 AND Zip matches
                            (
                              ( [address1] IS NOT NULL AND [add1] = @add2 ) OR  
                              ( [address2] IS NOT NULL AND [add2] = @add2 ) 
                            ) AND [zip] = @zip
                        )
                    )
                    GROUP BY customer_id, add1, add2, add3, address1, address2, address3, zip
                    HAVING COUNT(customer_id) > 1

        END 
        SET @cnt = @cnt + 1 
    END

如果我移除 GROUP BY 以及 HAVING 然后它返回每一行，因为#tmpcustomers只是工作#customeraddress字段的子集。这是有道理的。
如果我取消注解 [customer_id] != @customer_id 然后它只给我重复的记录信息，而不是源记录信息。如果b和c是a的副本，我只得到b和c，但我想要a，b和c。
我试着添加 GROUP BY 以及 HAVING 但后来我一点结果都没有。所有的记录应该是唯一的，因为它只是一个 per customer_id ，所以 GROUP BY 不应该是个问题。也许是 HAVING 是不对的。
我错过了什么？
我正在运行最新版本的mssqlserver和ssms