使用以下sql表:
表:财政
DateID | date | fiscal_year | fiscal_week
20170101 2017-01-01 00:00:00.0 2017 2017 WK 01
20170102 2017-01-02 00:00:00.0 2017 2017 WK 01
table:email_info
email_id | email_name | email_subjectline
123 New_Year_2017 Welcome the new year!
345 Reminder Don't forget
table:sent_info
email_id | sent_date
123 | 1/1/2017 8:58:39 PM
345 | 1/2/2017 6:33:39 AM
表:click_info
recipient | email_id | click_date
XYZ 123 1/7/2017 4:25:27 PM
ABC 123 1/5/2017 3:13:56 AM
CDF 345 1/6/2017 2:20:16 AM
ABC 345 1/14/2017 3:33:25 AM
显然每个表中有很多行。
电子邮件表之间的连接非常简单。
SELECT *
FROM email_info
JOIN sent_info
ON sent_info.email_id = email_info.email_id
JOIN click_info
ON click_info.email_id = email_info.email_id
我正在努力解决以下问题:
这是我正在寻找的输出(按财政周度= 2017年WK 01过滤):
email_id | email_name | sent_date | fiscal_week | Clicks
123 New_year_2017 1/1/2017 2017 WK 01 2
345 Reminder 1/2/2017 2017 WK 01 1
*请注意,click_info表示例中的最后一次点击未计算在内,因为它超出了发送日期后的7天。
** DateID是一个整数,sent_date和click_date是strings / varchar
答案 0 :(得分:2)
假设dateId是varchar而其他日期时间应该是
select a.email_id, a.email_name, date(b.sent_date), c.fiscal_week, count(d.click_date)
from email_info a
inner join fiscal c on str_to_date(c.dateID, '%Y%m%d') = date(b.sent_date)
inner join sent_info b on b.email_id = c.email_id
inner join click_info d on d.email_id = b.email_id
and date(d.click_date) between date(b.sent_date) and DATEADD(week,1,date(b.sent_date))
group by a.email_id, a.email_name, date(b.sent_date), c.fiscal_week
PS 如果您有其他格式可以正确转换并更改单件
,则会定义查询结构。