SQL GROUP BY 结果 - Salesforce Marketing Cloud答案

【问题标题】：SQL GROUP BY Results - Salesforce Marketing CloudSQL GROUP BY 结果 - Salesforce Marketing Cloud
【发布时间】：2021-04-12 18:04:09
【问题描述】：

我正在寻找所有重复记录，然后选择所有重复记录减去每组中最旧的记录，以便我可以删除重复记录并保留一条唯一记录。

当我运行这个查询时，我得到了我想要的结果。留给我一个唯一的电子邮件地址和最早的创建日期。

SELECT 
    EmailAddress,
    MIN(CreatedDate)
FROM [_ListSubscribers]
WHERE EmailAddress IN
    (
        SELECT EmailAddress
        FROM _ListSubscribers
        GROUP BY EmailAddress
        HAVING COUNT(EmailAddress) > 1
    )
GROUP BY EmailAddress

当我将 SubscriberKey 添加到查询中时，结果是双倍的！为什么会这样？我只想查看与我发现的在子查询中日期最早的 EmailAddress 相关联的 SubscriberKey。

SELECT 
    EmailAddress,
    SubscriberKey,
    MIN(CreatedDate)
FROM [_ListSubscribers]
WHERE EmailAddress IN
    (
        SELECT EmailAddress
        FROM _ListSubscribers
        GROUP BY EmailAddress
        HAVING COUNT(EmailAddress) > 1
    )
GROUP BY EmailAddress, SubscriberKey

【问题讨论】：

标签： sql group-by salesforce-marketing-cloud

【解决方案1】：

您获得了多条记录，因为您按SubscriberKey 分组。您需要匹配 EmailAddress 和 CreatedDate。尝试执行子查询并将其连接回原始表。

select 
[_ListSubscribers].EmailAddress,
[_ListSubscribers].SubscriberKey,
[_ListSubscribers].CreatedDate,
from
(
SELECT 
    EmailAddress,
    MIN(CreatedDate) as CreatedDate
    FROM [_ListSubscribers]
    GROUP BY EmailAddress, SubscriberKey
    Having count(EmailAddress)>1
) SubTbl
inner join
[_ListSubscribers] on
[_ListSubscribers].EmailAddress = SubTbl.EmailAddress
and
[_ListSubscribers].CreatedDate = SubTbl.CreatedDate

【讨论】：

【解决方案2】：

我正在寻找所有重复记录，然后选择所有重复记录减去每组中最旧的记录，以便我可以删除重复记录并保留一条唯一记录。

使用ROW_NUMBER():

select l.*
from (select l.*,
             row_number() over (partition by EmailAddress order by CreatedDate desc) as seqnum
      from _ListSubscribers l
     ) l
where seqnum > 1;

但是，如果您想删除除最新记录之外的所有记录，您可以使用：

delete from _ListSubscribers
    where CreatedDate < (select max(CreatedDate)
                         from _ListSubscribers l2
                         where l2.EmailAddress = _ListSubscribers.EmailAddress
                        );

如果您想要最旧的记录，您可以使用min() 而不是max() 翻转逻辑。

【讨论】：