Question

我的服务中有一个电子邮件递送系统（对于我的电子邮件订阅者）。我需要获取有关已发送邮件，点击次数和我的电子邮件的开放次数的统计信息。所有信息都存储在 PostgreSQL 数据库中。

我的订阅者列表（subscriber_collections）：

_id name
===========
 1  MySubscribers
 2  VIP Subscribers

所以我有一个表存储每个订阅者列表的电子邮件（emails，一对多关系）

_id  subscriber_collection_id  email_address
===========================================
 1               1              john.smith@email.com
 2               1              james3@email.com
 3               1              jay10@email.com
 4               2              second@email.com
 5               2              john.smith@email.com
 6               2              jay10@email.com

我还有一个表mail_activities，它存储有关不同电子邮件活动的信息（例如点击，打开，发送等等）。：

_id  message_id  message_subject  activity_date   email_address       type
============================================================================
 1      25           Hello msg!    2017-11-21   john.smith@email.com  Sends
 2      25           Hello msg!    2017-11-21   john.smith@email.com  Opens
 3      25           Hello msg!    2017-11-21   john.smith@email.com  Clicks
 4      25           Hello msg!    2017-11-21   jay10@email.com       Clicks
 5      25           Hello msg!    2017-11-21   jay10@email.com       Clicks
 6      16           Discount!     2017-11-21   second@email.com      Opens
 7      16           Discount!     2017-11-21   second@email.com      Opens
 8      16           Discount!     2017-11-21   james3@email.com      Opens

请注意，用户可以在同一封电子邮件中重复点击（或打开），信息将在数据库中。

我有一个SQL查询，它可以计算每个电子邮件和订阅者集合的不同活动：

SELECT
    MA.message_id
    ,MA.message_subject
    ,SC.name as subscriber_collection_name
    ,MA.date AS activity_date
    ,SUM(CASE WHEN MA.type='Sends' THEN 1 ELSE 0 END) as sends_count
    ,SUM(CASE WHEN MA.type='Opens' THEN 1 ELSE 0 END) as opens_count
    ,SUM(CASE WHEN MA.type='Clicks' THEN 1 ELSE 0 END) as clicks_count
FROM
    mail_activities MA
    INNER JOIN emails E ON MA.email_address = E.email_address
    INNER JOIN subscriber_collections SC ON E.subscriber_collection_id = SC._id

GROUP BY
    MA.message_id
    ,MA.message_subject
    ,SC._id
    ,SC.name
    ,MA.date                    
ORDER BY
    message_id

此查询正常。以下是此数据的结果：

问题：我需要获得唯一点击次数并打开每个email_address 的活动。我需要将此信息作为该查询中的其他列（unique_clicks和unique_opens）获取。例如，表Hello msg!中用户jay10@email.com对字母emails的两次点击应视为一次。如何在其他列中获取此信息？感谢...

P.S：抱歉我的英语不好......

Answer 1

如果我理解正确，您可以使用条件count(distinct)：

COUNT(DISTINCT CASE WHEN MA.type = 'Sends' THEN ma.message_id END) as sends_unique_count
COUNT(DISTINCT CASE WHEN MA.type = 'Opens' THEN ma.message_id END) as opens_unique_count
COUNT(DISTINCT CASE WHEN MA.type = 'Clicks' THEN ma.message_id END) as clicks_unique_count

我不太确定这如何适合您现有的查询。您希望每封电子邮件都有唯一的计数（似乎是说明），但您要按message_id进行汇总。

嗯。嗯。。。也许你想要不同的电子邮件地址：

COUNT(DISTINCT CASE WHEN MA.type = 'Sends' THEN ma.email_address END) as sends_unique_count
COUNT(DISTINCT CASE WHEN MA.type = 'Opens' THEN ma.email_address END) as opens_unique_count
COUNT(DISTINCT CASE WHEN MA.type = 'Clicks' THEN ma.email_address END) as clicks_unique_count

唯一点击次数（SQL查询）

1 个答案: