1

我想从日志中计算发送的邮件（主表：ex_deliverylog）及其收件人（详细信息表：ex_deliverylog）。下面的查询为 [session] 和 [recipients] 返回相同的值。简而言之，我无法对 [会话] 进行分组和计数。

    Select 
        deliveryaccount,
        DATEDIFF(d,deliverytime, getdate()) AS ago
        ,COUNT(ex_deliverylog.deliveryid) as session
        ,COUNT(ex_deliverylog_recipients.deliveryid) as recipients

        --,( select count(*) from ex_deliverylog_recipients where ex_deliverylog.deliveryid = ex_deliverylog_recipients.deliveryid )  

    from ex_deliverylog

    left join ex_deliverylog_recipients 
           on ex_deliverylog_recipients.deliveryid = ex_deliverylog.deliveryid


    group by
        deliveryaccount,
        DATEDIFF(d,deliverytime, getdate())


    order by ago, session desc

查询&结果：

在此处输入图像描述

表格和字段：

在此处输入图像描述

我如何计算两个会话及其总收件人？

4

4 回答 4

1

现在，您将获得相同的值，因为您的查询为 each 返回了一组行GROUP BY，并且对于每个COUNT()语句，所有这些行都已填充，因此您在计算这些字段的结果中收到相同的值。您需要计算这些 id 的唯一值。所以改变

Select 
    deliveryaccount,
    DATEDIFF(d,deliverytime, getdate()) AS ago
    ,COUNT(ex_deliverylog.deliveryid) as session
    ,COUNT(ex_deliverylog_recipients.deliveryid) as recipients

到

Select 
    deliveryaccount,
    DATEDIFF(d,deliverytime, getdate()) AS ago
    ,COUNT(distinct ex_deliverylog.deliveryid) as session
    ,COUNT(distinct ex_deliverylog_recipients.deliveryid) as recipients

如果这不能满足您的需求，我建议将您的会话数和收件人数分成单独的查询，您可以更好地控制这些查询。

于 2013-07-10T13:40:07.247 回答
1

如果我了解您要做什么，我认为您需要使用COUNT DISTINCT会话计数，而不仅仅是COUNT默认为COUNT ALL：

SELECT 
        deliveryaccount,
        DATEDIFF(d,deliverytime, getdate()) AS ago
        ,COUNT(DISTINCT ex_deliverylog.deliveryid) as session
        ,COUNT(ex_deliverylog_recipients.deliveryid) as recipients
FROM ex_deliverylog
LEFT JOIN ex_deliverylog_recipients 
       ON ex_deliverylog_recipients.deliveryid = ex_deliverylog.deliveryid
GROUP BY
    deliveryaccount,
    DATEDIFF(d,deliverytime, getdate())
ORDER BY ago, session desc

这样，会话计数将反映不同会话的数量，而接收者计数将反映不同接收者的数量。当既没有ALL也没有DISTINCT指定时，COUNT默认为ALL并且您会得到您正在经历的行为（即，两者的计数相同）。

于 2013-07-10T13:48:52.190 回答
1

我如何计算两个会话及其总收件人？
需要统计 ex_deliverylog_recipients.deliveryrecipientid 

SELECT 
        deliveryaccount
       ,DATEDIFF(d,deliverytime, getdate()) AS ago
       ,COUNT(DISTINCT ex_deliverylog.deliveryid) as session
       ,COUNT(DISTINCT ex_deliverylog_recipients.deliveryrecipientid) as recipients
FROM ex_deliverylog
LEFT JOIN ex_deliverylog_recipients 
       ON ex_deliverylog_recipients.deliveryid = ex_deliverylog.deliveryid
GROUP BY
    deliveryaccount,
    DATEDIFF(d,deliverytime, getdate())
ORDER BY ago, session desc

一个人可以多次收到同一封电子邮件。
EG 被分到两组，每个组都有人。
如果您需要唯一收件人的计数，那么 deliveryrecipientaddress （不是 deliveryrecipientid）： 

SELECT 
        deliveryaccount
       ,DATEDIFF(d,deliverytime, getdate()) AS ago
       ,COUNT(DISTINCT ex_deliverylog.deliveryid) as session
       ,COUNT(DISTINCT ex_deliverylog_recipients.deliveryrecipientaddress) as recipients
FROM ex_deliverylog
LEFT JOIN ex_deliverylog_recipients 
       ON ex_deliverylog_recipients.deliveryid = ex_deliverylog.deliveryid
GROUP BY
    deliveryaccount,
    DATEDIFF(d,deliverytime, getdate())
ORDER BY ago, session desc
于 2013-07-10T13:55:11.623 回答
0

SQL Server 有一个很好的 SQL 扩展来执行此操作。

WITH ROLLUP在您的子句后添加“ ” GROUP BY，它将在每个 GROUP BY 级别生成小计。

于 2013-07-10T15:15:51.043 回答