显示多条记录的第二个结果

Question

我有一个查询显示一位客户的第二个结果。我现在需要做的是在特定列表中显示每个客户的第二个结果（例如 20 个不同的客户 G/1931、G/1932、G/1933 等）。

我该怎么做？ Sql Server 2000 通过 SSMS 2005

1 个客户的当前查询是

SELECT TOP 1 link_to_client, call_ref
FROM
(
    SELECT TOP 2 link_to_client, call_ref
    FROM calls WITH (NOLOCK)
    WHERE link_to_client IN ('G/1931')
        AND call_type = 'PM'
    ORDER BY call_ref DESC
) x
ORDER BY call_ref

Answer 1

我认为最有效的方法是排除每个link_to_client 的最小值call_ref，然后取其最小值：

SELECT  calls.link_to_client, 
        MAX(calls.Call_Ref) call_ref
FROM    calls WITH (NOLOCK)
        LEFT JOIN 
        (   SELECT  link_to_client, MAX(Call_Ref) call_ref
            FROM    calls WITH (NOLOCK)
            WHERE   calls.call_type = 'PM'
            GROUP BY link_to_client
        ) MaxCalls
            ON MaxCalls.link_to_client = calls.link_to_client
            AND MaxCalls.Call_ref = calls.call_ref
WHERE   calls.call_type = 'PM'
AND     MaxCalls.link_to_Client IS NULL
GROUP BY calls.link_to_Client;

但是，如果您想扩展它以获取，例如每个客户的第 5 个，那么它就会开始变得混乱。在这种情况下，我会使用：

SELECT  calls.link_to_client, calls.call_ref
FROM    calls
WHERE   5 = (   SELECT  COUNT(*)
                FROM    calls c2
                WHERE   c2.link_to_Client = calls.link_to_Client
                AND     c2.call_ref <= calls.call_ref
            );

我的最后一条建议是升级到更新版本的 SQL-Server，您可以在其中使用ROW_NUMBER！

Examples on SQL Fiddle

感谢Nenad Zivkovic 提供的小提琴