cancel
Showing results for 
Search instead for 
Did you mean: 
Reply
RMDNA
Solution Sage
Solution Sage

Custom column - "best of these choices"

Hi all,

 

I have a table (see below) with the Absolute Percent Error of three different models. I'd like to create a custom column with the "best" (lowest APE) of those three.

 

In Excel, I have =MIN(Table[@[Model 1 APE]:[Model 3 APE]])

 

What's the PQ equivalent? I'm unsure how to specify a range of columns.

 

1.PNG

 

As usual, thanks!

1 ACCEPTED SOLUTION
SteveCampbell
Memorable Member
Memorable Member

So really, we should not use a calculated column in DAX. It will result in the column being recaulclated every time and slow performance.

You could write a measure, which would not store a value for every row. As you asked for a column, instead, we should do this in the query editor. This will mean it is only calculated on a data refresh, and stored in the model.

 

In the query editor, go to the query, and select "Add Custom Column".

 

The code you can use is:

 

 

each if [Model 1 APE] = [Model 2 APE]then "N/A"

else if [Model 1 APE]< [Model 2 APE] then "Model 1 APE"
else if [Model 2 APE]< [Model 1 APE] then "Model 2 APE"
else "N/A"

Here, the second else if statement is a bit redundant, but I left it in so that you can add more logic if you wanted, to compare more models. You can add as many "else if" statements to use any logic you wish.

 



Did I answer your question? Mark my post as a solution! Proud to be a Super User!


Connect with me!
Stay up to date on  
Read my blogs on  



View solution in original post

4 REPLIES 4
SteveCampbell
Memorable Member
Memorable Member

So really, we should not use a calculated column in DAX. It will result in the column being recaulclated every time and slow performance.

You could write a measure, which would not store a value for every row. As you asked for a column, instead, we should do this in the query editor. This will mean it is only calculated on a data refresh, and stored in the model.

 

In the query editor, go to the query, and select "Add Custom Column".

 

The code you can use is:

 

 

each if [Model 1 APE] = [Model 2 APE]then "N/A"

else if [Model 1 APE]< [Model 2 APE] then "Model 1 APE"
else if [Model 2 APE]< [Model 1 APE] then "Model 2 APE"
else "N/A"

Here, the second else if statement is a bit redundant, but I left it in so that you can add more logic if you wanted, to compare more models. You can add as many "else if" statements to use any logic you wish.

 



Did I answer your question? Mark my post as a solution! Proud to be a Super User!


Connect with me!
Stay up to date on  
Read my blogs on  



Wow. Somehow, after almost three years of using Power BI, I entirely forgot about the Conditional Column button. Thanks for the support!

SteveCampbell
Memorable Member
Memorable Member

 

You can try:

New Column = 
_min_amnt = min([Model 1 APE],min([Model 2 APE],[Model 3 APE])

RETURN 

SWITCH( _min_amnt, 
  [Model 1 APE], "Model 1 APE",
  [Model 2 APE], "Model 2 APE",
  [Model 3 APE], "Model 3 APE",
  "None"
)

 

Although, this will not work in case of ties, and pick the first. What would be the desired solution if all 0, for example?

 



Did I answer your question? Mark my post as a solution! Proud to be a Super User!


Connect with me!
Stay up to date on  
Read my blogs on  



@SteveCampbell,

 

How about with two columns (Model 1 and 2)? In the case of a tie/all zero, display "N/A."

Helpful resources

Announcements
May 23 2022 epsiode 5 without aka link.jpg

The Power BI Community Show

Welcome to the Power BI Community Show! Jeroen ter Heerdt talks about the importance of Data Modeling.

charticulator_carousel_with_text (1).png

Charticulator Design Challenge

Put your data visualization and design skills to the test! This exciting challenge is happening now through May 31st!

Power BI Dev Camp Session 22 with aka link 768x460.jpg

Check it out!

Mark your calendars and join us on Thursday, May 26 at 11a PDT for a great session with Ted Pattison!

Power BI Release May 2022 768x460.png

Check it out!

Click here to read more about the May 2022 updates!