聚合功率 BI SUM 的问题

Question

I'm trying to replicate a business calculation that exists in my environment on PowerBI, but I'm not getting what I expect as a result.

I have 2 tables that I made as an example and replicate the real life problem.

Table Sales:

dateSold      Sector    Idsale  NPS
01/12/2021    A         1       good
01/11/2021    A         2       bad
01/11/2021    A         3       good
01/11/2021    B         4       bad
01/10/2021    A         5       bad
01/10/2021    B         6       good
01/10/2021    B         7       good
01/09/2021    A         8       good
01/09/2021    A         9       good
01/09/2021    A         10      good

And Weights table:

SECTOR  Weight
A       0,7
B       0,3

I created some fields to do my math and they do what I expect, but the last field doesnt. I did a COUNT for GOOD and BAD and total values on the field NPS using this method:

QuantityGOOD = 
CALCULATE(
    COUNTA(Sales[Idsale])
    , DATESINPERIOD(Sales[dateSold], SELECTEDVALUE(Sales[dateSold]), -3, MONTH)
    , Sales[NPS] == "good"
)

Then Created a field with the math using my fields:

NPSCalculated = 
[QuantityGOOD] / [QuantityTotal] - [QuantityBad] / [QuantityTotal] 

Then in the last step I have to multiply this value by the weights in the other table, and I tryed this:

FinalValue = 
CALCULATE(
SUMX(
    SUMMARIZE(Sales ,Sales[dateSold], weights[Sector]),
    [NPSCalculated] * 
    CALCULATE(
        AVERAGE(weights[weight]))))

But this code cuts out the values of the sector A that exists in the previous months and doesnt exist in the target month. It returns the sum without the part missing.

SaleDate    Sector  bad   good  total   NPS     weight  FinalValue
2021-12     A       2     2     4       0       0,7     0
2021-12     B       1     2     3       0,333   0,3     0,1

I need to get the results of the column FinalValue as above (did it in excel for 2021-12 only), but instead I get only the fist line on PBI. Can you guys help me?

Answer 1

CALCULATE(
    SUMX(
        CALCULATETABLE(
             SUMMARIZE(Sales ,Sales[dateSold], weights[Sector])
             ,DATESINPERIOD(Sales[dateSold], SELECTEDVALUE(Sales[dateSold]), -3, MONTH)
        )
        ,[NPSCalculated] * 
        CALCULATE(
            AVERAGE(weights[weight])
        )                            
    )
)

Answer 2

1. Check the following steps are followed Click on File-->Get data-->and import the required data-->then load the data into PowerBI Loading The Data

Make sure that the dataset is loaded perfectly.

2. Next, you need to create a new date table for this visual via Calendar = CALENDAR(MIN('Sales'[dateSold]), MAX('Sales'[dateSold])) Then add a calculated column named month like the below.

calendar table creation

3. In the table visual, replace [dateSold] with [Month], which will let the table visual get a Cartesian product of [month] and [SECTOR].

4. And change the parameter in your measure DATESINPERIOD(). Replace the dateSold column with the new date column. DATESINPERIOD('Sales'[dateSold], MIN('Calendar'[Date]),-3,MONTH)) dates in period

I need to get the results of the column FinalValue as above (did it in excel for 2021-12 only), but instead I get only the first line on PBI. Can you guys help me?

Then in the last step, I have to multiply this value by the weights in the other table, and I tried this:

FinalValue = CALCULATE( SUMX( SUMMARIZE(Sales,Sales[dateSold], weights[Sector]), [NPSCalculated] * CALCULATE( AVERAGE(weights[weight]))))

Instead of this if we use this code then the expected output will be shown as the result FinalValue = VAR _s = ADDCOLUMNS ( CROSSJOIN ( VALUES ( 'Calendar'[Month] ), 'Weight' ), "final", [Weight] * [NPSCalculated] ) RETURN SUMX ( _s, [final] ) Result