Calculate difference between rows

[Martin_Bruwer]

Hello All,

I'm trying to calculate the number of days between two appointments where the first appointment could be one of two types. in the below table it would be the days between the earlier of appointment type one or two and appointment three.

Patient ID	Appointment Type	Appintment Date
112233	Type 1	01/10/2017
112233	Type 2	01/11/2017
112233	Type 3	01/01/2018
223344	Type 1	01/10/2017
223344	Type 2	01/09/2017
223344	Type 3	01/01/2018

For for example patient 112233, would be days between appointment type 1 (being the earlier of appointment 1 and 2) and appointment type 3, so 61 days.

I am thinking a calculated column which returns the earliest date of the two filtered by patient ID and then just a datediff between the two dates on the row with appointment type 3.

Any thoughts?

Martin

[Anonymous]

Depends on if you need the values as a column, or if a measure would suffice (measures are preferred from a performance standpoint).

Try this measure:

[Measure] =
VAR CurrentPatient =
    VALUES ( TableName[Patient ID] )
VAR Appointment3Date =
    CALCULATE (
        MIN ( TableName[Appointment Date] ),
        TableName[Appointment Type] = "Type 3",
        TableName[Patient ID] IN CurrentPatient
    )
VAR EarliestAppointmentDate =
    CALCULATE (
        MIN ( TableName[Appointment Date] ),
        TableName[Appointment Type] IN { "Type 1", "Type 2" },
        TableName[Patient ID] IN CurrentPatient
    )
RETURN
    Appointment3Date - EarliestAppointmentDate

Variables make it easy to call out what each part of the calculation is doing.

[Martin_Bruwer]

Thank you @Zubair_Muhammad and @Anonymous

Chris,

I used your measure but made a slight change to the return,

Changed your to

  DATEDIFF(EarliestAppointmentDate,AppointmentDate,day)

Giving me days as opposed to a date formate.

It appears to be working, time to do testing.

Thank you so much.

[Martin_Bruwer]

@Anonymous

Hi Again,

Immidiately ran into two issues.

1. Because of the nature of the business it is possible to have a patient who has appointment type 3 prior to appointment types 1 or 2. Using Datediff I get the error that the start date cannot be greater than the end date.  I could not use datediff but I need to have a result in days rather than a date.
2. The result is also needed to be and Median and Mean over many hundreds of patients, with the measure I'm not sure where to put the function in.

Thank you again.

Martin

[Zubair_Muhammad · Accepted solution]

@Martin_Bruwer

Measure =
VAR EarlierAppointment =
    CALCULATE (
        MIN ( TableName[Appintment Date] ),
        FILTER (
            ALLEXCEPT ( TableName, TableName[Patient ID] ),
            TableName[Appointment Type] = "Type 1"
                || TableName[Appointment Type] = "Type 2"
        )
    )
VAR Type3Date =
    CALCULATE (
        FIRSTNONBLANK ( TableName[Appintment Date], 1 ),
        FILTER (
            ALLEXCEPT ( TableName, TableName[Patient ID] ),
            TableName[Appointment Type] = "Type 3"
        )
    )
VAR Start_date =
    MIN ( EarlierAppointment, Type3Date )
VAR End_date =
    MAX ( EarlierAppointment, Type3Date )
RETURN
    DATEDIFF ( Start_date, End_date, DAY )

[Martin_Bruwer]

Thank you very much, I believe I have a solution now. Trial and error.

[Zubair_Muhammad]

Hi @Martin_Bruwer

Try this calculated column

Column =
VAR EarlierAppointment =
    CALCULATE (
        MIN ( TableName[Appintment Date] ),
        FILTER (
            ALLEXCEPT ( TableName, TableName[Patient ID] ),
            TableName[Appointment Type] = "Type 1"
                || TableName[Appointment Type] = "Type 2"
        )
    )
VAR Type3Date =
    CALCULATE (
        FIRSTNONBLANK ( TableName[Appintment Date], 1 ),
        FILTER (
            ALLEXCEPT ( TableName, TableName[Patient ID] ),
            TableName[Appointment Type] = "Type 3"
        )
    )
RETURN
    DATEDIFF ( EarlierAppointment, Type3Date, DAY )

[Zubair_Muhammad]

@Martin_Bruwer

Hopefully This formula will work as a MEASURE as well