Average duration between two dates is not calculated

[JoSwinnen]

Hi,

I have a dataset with multiple lines for one student. For each year he is stuying at the university a line is created with his registrationdate of that year and his deregistrationdate of that year. I want to calculate how long students are studying on average. 

I've tried this one, based on https://community.powerbi.com/t5/Desktop/Calculate-Differance-Between-two-dates-grouping-by-another-... 

Study duration=

    VAR Startdate = calculate(min('vw_powerBI'[registrationdate]),ALLEXCEPT(vw_powerBI,vw_powerBI[studentnumber]))

    VAR Enddate = calculate(max('vw_powerBI'[deregistrationdate]),ALLEXCEPT(vw_powerBI,vw_powerBI[studentnumber]))

    RETURN

    AVERAGEX(vw_powerBI,DATEDIFF(Startdate, Enddate,DAY))

But it is not giving the appropriate result. 

Can anyone help? 

Kind regards,

Jo 

[JoSwinnen]

Thank you for your help! I think I'm almost there, but I'm working on another project today. I will get back to it!

[tamerj1 · Accepted solution]

 
actually when I saw your reply I had a 2nd look at the solution and the results. 
In your report you are slicing by 'vw_powerBI'[stamnummer] which means that the filter context contains the table of each student and there is no need to CALCULATETABLE - ALLEXCEPT. Therefore, the VALUES ( 'vw_powerBI'[Graduated?] ) should contain all the "Yes" and "No" values for each "stamnummer". The limitation I set is: IF at least one "Yes" is available for the current student then do the calculation otherwise return blank. I see no reason why this shouldn't work.

However this condition need to be applied inside the SUMX as well in order to limit the total average (In the column total cell) to only the students who have "Yes" in their Graduate status. 

Study duration =
IF (
    "Yes" IN VALUES ( 'vw_powerBI'[Graduated?] ),
    AVERAGEX (
        VALUES ( 'vw_powerBI'[stamnummer] ),
        IF (
            "Yes" IN CALCULATETABLE ( VALUES ( 'vw_powerBI'[Graduated?] ) ),
            DATEDIFF (
                CALCULATE ( MIN ( 'vw_powerBI'[registrationdate] ) ),
                CALCULATE ( MAX ( 'vw_powerBI'[deregistrationdate] ) ),
                DAY
            )
        )
    )
)

@JoSwinnen

[JoSwinnen]

My apologies for the late reply but now it works, thanks a lot!!! 

[amitchandak]

@JoSwinnen , Try like

Study duration=
VAR Startdate = calculate(min('vw_powerBI'[registrationdate]),ALLEXCEPT(vw_powerBI,vw_powerBI[studentnumber]))
VAR Enddate = calculate(max('vw_powerBI'[deregistrationdate]),ALLEXCEPT(vw_powerBI,vw_powerBI[studentnumber]))
RETURN
AVERAGEX(values(vw_powerBI,vw_powerBI[studentnumber]),calculate(DATEDIFF(Startdate, Enddate,DAY)))

[JoSwinnen]

Thanks for this. Now it gives for each student the correct duration but the average is not correct. 

And there is another difficulty that I'm realizing. I only want to calculate the study duration if this student has graduated. 

So there are different rows for one student and in the last year there will be 'yes' in the column 'graduated?' but in the previous rows there will be 'no'. 

Any advice? 

Thanks,

Jo 

[tamerj1]

Hi @JoSwinnen 
Please try the following measure. Also I would really appreciate if you share more details about your visual and what columns are you slicing by in this visual.

Study duration =
AVERAGEX (
    VALUES ( 'vw_powerBI'[studentnumber] ),
    CALCULATE (
        DATEDIFF (
            MIN ( 'vw_powerBI'[registrationdate] ),
            MAX ( 'vw_powerBI'[deregistrationdate] ),
            DAY
        )
    )
)

[JoSwinnen]

Thanks. This one gives a better result. 

The dataset looks like this:

Studentnumber       RegistrationDate      DeregistrationDate     Graduated?

00001                            15/09/2019               30/06/2020                No

00001                            15/09/2020               30/06/2021                No

00001                            15/09/2021               30/06/2022                Yes

00002                             15/09/2020              30/06/2021                  No

00002                             15/09/2021              30/06/2022                  No

00003                              ...

So the formula above may only work when there is one row with a yes for that student. Students who tried but did not graduated (only no's) cant be included in the calculation. 

Thanks,

Jo 

[tamerj1]

@JoSwinnen 
Please try

Study duration =
VAR Graduated =
    CALCULATETABLE (
        VALUES ( 'vw_powerBI'[Graduated?] ),
        ALLEXCEPT ( 'vw_powerBI', 'vw_powerBI'[Studentnumber] )
    )
RETURN
    AVERAGEX (
        VALUES ( 'vw_powerBI'[studentnumber] ),
        CALCULATE (
            IF (
                "Yes" IN Graduated,
                DATEDIFF (
                    MIN ( 'vw_powerBI'[registrationdate] ),
                    MAX ( 'vw_powerBI'[deregistrationdate] ),
                    DAY
                )
            )
        )
    )

Or 

Study duration =
VAR CurrentStudentTable =
    CALCULATETABLE (
        'vw_powerBI',
        ALLEXCEPT ( 'vw_powerBI', 'vw_powerBI'[Studentnumber] )
    )
VAR FiteredTable =
    FILTER ( CurrentStudentTable, 'vw_powerBI'[Graduated?] = "Yes" )
RETURN
    AVERAGEX (
        FiteredTable,
        DATEDIFF (
            MIN ( 'vw_powerBI'[registrationdate] ),
            MAX ( 'vw_powerBI'[deregistrationdate] ),
            DAY
        )
    )

[JoSwinnen]

Hi,

I've tried both ways, but it all gives the same result. The study duration for each student graduated or not... 

Thanks,

Jo 

[tamerj1]

@JoSwinnen 

Please try

Study duration =
VAR Graduated =
    CALCULATETABLE (
        VALUES ( 'vw_powerBI'[Graduated?] ),
        ALLEXCEPT ( 'vw_powerBI', 'vw_powerBI'[Studentnumber] )
    )
RETURN
    IF (
        "Yes" IN Graduated,
        AVERAGEX (
            VALUES ( 'vw_powerBI'[studentnumber] ),
            CALCULATE (
                DATEDIFF (
                    MIN ( 'vw_powerBI'[registrationdate] ),
                    MAX ( 'vw_powerBI'[deregistrationdate] ),
                    DAY
                )
            )
        )
    )

[JoSwinnen]

Thank you for trying, but it still gives the same result. 

[tamerj1]

@JoSwinnen 

Would you please share a screenshot of the code you've been using

[tamerj1]

@JoSwinnen 

I need to have a look at the data. Maybe we can connect tomorrow if you wish.