Active users last x days

user09335
user09335 Member
in Magic ETL

I have the following dataset ( see the attached screenshot). It shows the userID active on a given date and the session time. I want to determine which users were active in the current week, t-1 week, t-2 week etc. As an output I want a table with the following columns: UserID, active this week column ( 0 or 1 value), active last week (0 or 1 value) etc. How can I transform my current dataset in ETL or create calculated fields to create such an output. Any help would be greatly appreciated!

 

 

 

Best Answer

  • BlueRooster
    BlueRooster Domo Employee
    Answer ✓

    If you were to use a SQL transform it would look something like this:

    SELECT `User ID`, 
        CASE WHEN SUM(CASE WHEN Week(`Date`) = Week(Curdate()) 
                  AND Year(`Date`) = Year(Curdate()) 
             THEN 1 
             ELSE 0 
        END) >= 1 THEN 1 ELSE 0 END AS 'Current Week',
        CASE WHEN SUM(CASE WHEN Week(`Date`) = Week(DATE_SUB(Curdate(), INTERVAL 1 Week)) 
                  AND Year(`Date`) = Year(DATE_SUB(Curdate(), INTERVAL 1 Week)) 
             THEN 1 
             ELSE 0 
        END) >= 1 THEN 1 ELSE 0 END AS 'Last Week'
    FROM dataset
    GROUP BY `User ID`

    That should give you the Userid, CurrentWeek and LastWeek columns. If you need additional, just modify the Date_Sub Intervals for the Last Week example.

     

    Hope this helps,

    ValiantSpur

     

    **Please mark "Accept as Solution" if this post solves your problem
    **Say "Thanks" by clicking the "heart" in the post that helped you.

Answers

  • BlueRooster
    BlueRooster Domo Employee
    Answer ✓

    If you were to use a SQL transform it would look something like this:

    SELECT `User ID`, 
        CASE WHEN SUM(CASE WHEN Week(`Date`) = Week(Curdate()) 
                  AND Year(`Date`) = Year(Curdate()) 
             THEN 1 
             ELSE 0 
        END) >= 1 THEN 1 ELSE 0 END AS 'Current Week',
        CASE WHEN SUM(CASE WHEN Week(`Date`) = Week(DATE_SUB(Curdate(), INTERVAL 1 Week)) 
                  AND Year(`Date`) = Year(DATE_SUB(Curdate(), INTERVAL 1 Week)) 
             THEN 1 
             ELSE 0 
        END) >= 1 THEN 1 ELSE 0 END AS 'Last Week'
    FROM dataset
    GROUP BY `User ID`

    That should give you the Userid, CurrentWeek and LastWeek columns. If you need additional, just modify the Date_Sub Intervals for the Last Week example.

     

    Hope this helps,

    ValiantSpur

     

    **Please mark "Accept as Solution" if this post solves your problem
    **Say "Thanks" by clicking the "heart" in the post that helped you.

  • user09335
    user09335 Member

    Perfect! It worked. Thank you!

Categories