Calculate an average with date filter

[Pedro77000]

Hi Everyone,

I have a problem of averaging according to a context. I have a data model with a fact table and a "Date" dimension table

Indeed, I want to calculate the average based on the month selected to calculate the average over the last 12 months (12 rolling months). Look at the following table:

13/10/2019	6
21/10/2019	3
01/09/2019	15
14/09/2019	1
20/11/2019	3
27/11/2019	3
29/11/2019	6
30/11/2019	1

Par exemple :
- For example, the average I want to have when I select September  is: (15 + 1)/ 1= 16

- For example, the average I want to have when I select October is: (16 + 9)/2 = 12,5

- For example, the average I want to have when I select November  is: (16 + 9+13)/3 = 12,66

For information, the relationship between my fact table and the "Date" dimension is based on a "Date" field.

Thank you very much in advance !!

[v-frfei-msft · Accepted solution]

Hi @Pedro77000 ,

We can create a measure as below.

Measure =
VAR SUMA =
    CALCULATE (
        SUM ( 'Table'[value] ),
        FILTER ( ALL ( 'Table' ), 'Table'[date] <= MAX ( 'date'[Date] ) )
    )
VAR COUNTM =
    CALCULATE (
        DISTINCTCOUNT ( 'Table'[YearMOnth] ),
        FILTER ( ALL ( 'Table' ), 'Table'[date] <= MAX ( 'date'[Date] ) )
    )
RETURN
    DIVIDE ( SUMA, COUNTM )

Also you can find the pbix as attached.

[v-frfei-msft · Accepted solution]

Hi @Pedro77000 ,

We can create a measure as below.

Measure =
VAR SUMA =
    CALCULATE (
        SUM ( 'Table'[value] ),
        FILTER ( ALL ( 'Table' ), 'Table'[date] <= MAX ( 'date'[Date] ) )
    )
VAR COUNTM =
    CALCULATE (
        DISTINCTCOUNT ( 'Table'[YearMOnth] ),
        FILTER ( ALL ( 'Table' ), 'Table'[date] <= MAX ( 'date'[Date] ) )
    )
RETURN
    DIVIDE ( SUMA, COUNTM )

Also you can find the pbix as attached.

[Pedro77000]

Good morning v-frfei-msft !!

Thank you very much for this reply.

The solution works well. But this average I cannot have it by region, I believe because of the filter ALL.

The average for the total is good but when I split by region, it's the same overall average that appears in front of each region.

 

[v-frfei-msft · Accepted solution]

Hi @Pedro77000 ,

To use ALLEXCEPT instead of ALL should work.

FILTER ( ALLEXCEPT ( 'Table','Table'[region] ), 'Table'[date] <= MAX ( 'date'[Date] ) )

If it doesn't meet your requirement,  Kindly share your sample data and excepted result to me if you don't have any Confidential Information. Please upload your files to One Drive and share the link here.

[Pedro77000]

Thank you so much Franck ! that works well ! even if the disadvantage of ALLEXCEPT is having to enter the fields of all dimension tables.

[Anonymous]

Hi @Pedro77000 

Try changing ALL to ALLSELECTED and then it should carry your filters through.

---
Please hit the "Accept as Solution" button if my post answered your question! If my post was helpful please consider giving it a "Thumbs Up."