How to use this CASE statement to DAX or to a Custom Column?

[BBIUser]

Hi,

Having trouble to get the result out of these CASE statements. Sample data is attached below.

1)

sum(case when Column1='Yes' or Column2='Yes')then 1 else 0 end) TotalValBranches

2) 

sum(case when (Column1 = 'Yes' or Column2 = 'Yes') and Current=0 and Expected=0 then 1 else 0 end) TotalAttBranches

3) 

sum(case when Coulmn3 = 'Yes' then 1 else 0 end) TotalChBranches

Ho do I use these CASE statements as a DAX? I will be counting the Total row count and might use it as a card visual.

Also, is it better to create a calculated Measure from these CASE statements or a Custom Column and then use that as a DAX?

 Thanks in Advance!

[MFelix · Accepted solution]

Hi @BBIUser,

You are correct this ask for a table if want to apply any filters you should wrap your column in a ALLSELECTED if you want an overall result the use the ALL function.

In the case the measure would be something like this:

TotalValBranches = calculate (countrows(ALLSELECTED(Table[Current]); Table[Column1] ='Yes' || Table[Column2] ='Yes')

This should work. Then just divide 1 measure by the other to get the percentage.

Regarding the columns vs measure best practices on DAX says thst if you can calculate it making a measure don't use columns.

Columns add size to your model and decrease performance since you are adding values to your model, assuming you table as 100.000 rows adding one column is adding 100.000 additional data cells.
Measures are only calculated when they are called and also store on cache for next uses.

Regards
MFelix

[MFelix]

Hi @BBIUser,

In DAX you need to define the filter to make this calculations work correcly (in this case) in other you may need to use the Switch or calculate function.

Try the formulas below:

TotalValBranches = calculate (countrows(Table[Current] ; Table[Column1] ='Yes' || Table[Column2] ='Yes')


TotalAttBranches = calculate (countrows(Table[Current] ; Table[Column1] ='Yes' || Table[Column2] ='Yes';Table[Current] = 0 ; Table[Expected] = 0 )

TotalChBranches = calculate (countrows(Table[Current] ; Table[Column3] ='Yes')

These 3 measures should give expected result.

Regards,

MFelix

[BBIUser]

Appreciate your reply @MFelix. 

Does the CALCULATE(COUNTROWS(.. DAX accept the Column name after the COUNTROWS? I was getting errors and the expression was expecting a Table Name.

I tried creating an 'Custom Columns' for all the CASE statements. Below is an example in order.

TotalValBranches

if ([Column1] = "Yes" or [Column2] = " Yes") then 1 else 0

 TotalAttBranches

if (([Column1] = "Yes" or [Column2]= "Yes") and [Current]=0 and [Expected]=0) then 1 else 0

TotalChBranches

if (Column3 = "Yes") then 1 else 0

I am not confident creating a custom column is correct or not, but I am still having trouble defining DAX from these custom columns too.

My final result should show TotalValBranches/TotalAttBranches values that will be displayed in a card or multi-row card as (numerator and denominator). Finally the % from these Numtor and Denomtor.

[v-lili6-msft · Accepted solution]

hi, @BBIUser

Your formulas are M language in the Edit Queries.

For creating the calculate column by DAX 

You could click Modeling ->New Column

Then use these formulae to create three column

TotalValBranches = IF([Column1] = "Yes" || [Column2] = " Yes", 1,  0)

TotalAttBranches = if (([Column1] = "Yes" || [Column2]= "Yes") && [Current]=0 && [Expected]=0, 1 , 0)

TotalChBranches = IF([Column3] = "Yes",  1 , 0)

Comparison

And then add a measure for TotalValBranches/TotalAttBranches by DIVIDE Function

TotalValBranches/TotalAttBranches =
DIVIDE (
    CALCULATE ( SUM ( 'Table'[TotalValBranches] ) ),
    CALCULATE ( SUM ( 'Table'[TotalAttBranches] ) ),
    0
)

and drag it into card visual.

also, you could just create a measure like below:

Measure = 
DIVIDE (
    CALCULATE (
        COUNTROWS ( 'Table' ),
        FILTER ( 'Table', [Column1] = "Yes" || [Column2] = " Yes" )
    ),
    CALCULATE (
        COUNTROWS ( 'Table' ),
        FILTER (
            'Table',
            ( [Column1] = "Yes"
                || [Column2] = "Yes" )
                && [Current] = 0
                && [Expected] = 0
        )
    ),
    0
)

Columns add size to your model and decrease performance since you are adding values to your model and they are static.

Measures are only calculated when they are called and also store on cache for next uses and they are dynamic.

here is my pbix, please try it.

https://www.dropbox.com/s/sl6bmsjw2302ihg/How%20to%20use%20this%20CASE%20statement%20to%20DAX%20or%2...

Best Regards,

Lin

[BBIUser]

Thanks for the insights! Very informative @MFelix and @v-lili6-msft. 

It worked.

You guys rock!

[MFelix · Accepted solution]

Hi @BBIUser,

You are correct this ask for a table if want to apply any filters you should wrap your column in a ALLSELECTED if you want an overall result the use the ALL function.

In the case the measure would be something like this:

TotalValBranches = calculate (countrows(ALLSELECTED(Table[Current]); Table[Column1] ='Yes' || Table[Column2] ='Yes')

This should work. Then just divide 1 measure by the other to get the percentage.

Regarding the columns vs measure best practices on DAX says thst if you can calculate it making a measure don't use columns.

Columns add size to your model and decrease performance since you are adding values to your model, assuming you table as 100.000 rows adding one column is adding 100.000 additional data cells.
Measures are only calculated when they are called and also store on cache for next uses.

Regards
MFelix