AVERAGEX with extra rows

[Noeleke1301]

Hi all,

I'm trying to calculate an average of a percentage. However, based on a measure that counts specific events, I want to include extra rows to be averaged.

For example, the following table:

1	30%
2	20%
3	50%
4	100%

The average of above is 50%

If the measure counts 2 occurences, I want to add 0% for each occurence. So like this:

1	30%
2	20%
3	50%
4	100%
5	0%
6	0%

Now the average of the table is 33.33%

My current DAX seems to take the extra rows into account when calculating (the average changes when I change 0 to 1 for example), but never returns the correct average percentage.

Basically I select 2 column from table1 (one dummy + one with the percentage rows) and UNION it on a self-generated table that generates the amount of rows that the measure counts, then adds a 0% for each row.

Average with extra rows = 
var _count = [count measure]
return

AVERAGEX( 
    UNION( 
        SELECTCOLUMNS( Table1 , "Col1" , Table1[col1] ,"Pct" , AVERAGE( Table1[Pct] ) ) , 
    ADDCOLUMNS( GENERATESERIES( 0 , _count ) , "Pct" , 0 ) )  , 
    [Pct] )

Please help, thanks

[v-tangjie-msft · Accepted solution]

Hi @Noeleke1301 ,

We can create this measure.

Average with extra rows = 
var _count = [count measure]
return

AVERAGEX( 
    UNION( 
        SELECTCOLUMNS( Table1 , "Col1" , Table1[col1] ,"Pct" , Table1[Pct] )  , 
    ADDCOLUMNS( GENERATESERIES( 0 , _count-1 ) , "Pct" , 0 ) )  , 
    [Pct] )

If the measure counts 2 occurences, the average of the table is 33.33%.

If the above one can't help you get the desired result, please provide some sample data in your tables (exclude sensitive data) with Text format and your expected result with backend logic and special examples. It is better if you can share a simplified pbix file. Thank you.

Best Regards,

Neeko Tang

If this post  helps, then please consider Accept it as the solution  to help the other members find it more quickly. 

[Noeleke1301]

Awesome, thanks so much! Those tweaks you did made it work beautifully, exactly what I wanted.