Subtotal column based on time and text column

[prakhar]

I have a data that looks like this

Date | Time | Text | Sum
1-1-2001 | 12:23:34 | a| 2
1-1-2001 | 12:23:34 | b| 3
1-1-2001 | 12:25:34 | a| 2
1-1-2001 | 12:26:34 | b| 2
1-1-2001 | 13:23:34 | a| 1
1-1-2001 | 13:24:34 | a| 2

I need a result with something like this
Text | Sum
a | 4
b | 5
a | 3

or even just the sum column.

The first and third row are grouped because they are within a 5 minute range. Similarly with 2, 4  and 5,6. My end goal is to determine the frequency of each sum value in the data. It is a large dataset with multiple text values and date values.

[v-yulgu-msft]

Hi @prakhar,

Based on my understanding, you need to sum the column [Sum] for each [Text] type every 5 minutes, right?

In this scenario, please create some calculated columns, referring to below DAX formulas:

Min Time for each text =
CALCULATE (
    MIN ( 'SubTotal Column'[Time] ),
    ALLEXCEPT (
        'SubTotal Column',
        'SubTotal Column'[Date],
        'SubTotal Column'[Text]
    )
)

Difference Minute =
HOUR ( 'SubTotal Column'[Min Time for each text] - 'SubTotal Column'[Time] )
    * 60
    + MINUTE ( 'SubTotal Column'[Time] - 'SubTotal Column'[Min Time for each text] )
    + SECOND ( 'SubTotal Column'[Time] - 'SubTotal Column'[Min Time for each text] )
        / 60

Interval(every 5 minute) =
INT ( 'SubTotal Column'[Difference Minute] / 5 )

Total Sum =
CALCULATE (
    SUM ( 'SubTotal Column'[Sum] ),
    ALLEXCEPT (
        'SubTotal Column',
        'SubTotal Column'[Date],
        'SubTotal Column'[Text],
        'SubTotal Column'[Interval(every 5 minute)]
    )
)

Data view

Table visual

Best regards,
Yuliana Gu

[prakhar]

I was able to create what I needed using your idea of adjusted time. I did this:

Adjusted Time = CALCULATE(
                MIN('items'[Time]),
                ALL('items'[Date])   
                ALL('items'),
                HOUR('items'[Time]-EARLIER('items'[Time]))*60
                +MINUTE('items'[Time]-EARLIER('items'[Time]))<=5)

[prakhar]

Hi @v-yulgu-msft,

Thanks for the reply. That is not exactly what I need.

I was also thinking of a similar solution. But with this solution, the time is divided into fixed 5 minute intervals. I don't want to divide the time. I want to put in an if condition that is based on whether the rows that need to be grouped are within 5 minutes of each other or not. eg. by dividing into fixed intervals, 12:20 and 12:24 text are grouped and  12:29 and 12:31 won't be grouped. I need both the groupings to occur

What I think would be better is to compare rows and if two rows with same Text are within 5 minutes of each other, then add their sum values. I am not sure how this could be implemented in Power BI. I saw EARLIER function used for such cases but haven't been able to write query using them or even understand them properly.

Regards
Prakhar