Variables in Power Query, Error Handling

[Anonymous]

Hello,

Apologies for my naivete and inexperience regarding PowerQuery and DAX, in advance.  I need some help understanding something.  

I've reached out twice now for solutions to some computational issues I've been dealing with in automating reports through the use of PowerQuery and PowerPivot.  In both cases I received brilliant responses which included code that utilized what appeared to be variable declarations to accomplish the challenges.  I understand enough to see how they should work to solve my situations, but in both cases I receive the same error (shown below), which makes me think I'm misunderstanding something about the functionality of DAX (?) in PowerQuery.

https://community.powerbi.com/t5/Power-Query/Linking-TimeStamp-to-Value-between-Dates-in-Seperate-Ta...

https://community.powerbi.com/t5/Power-Query/DateDiff-between-rows-of-Status-changes/m-p/764456#M255...

If someone could help me to better understand this, what I'm doing wrong or misunderstanding about coding in DAX, I would be most grateful!

Sincerely,

Kristopher

[ImkeF · Accepted solution]

Hi @Anonymous ,

you cannot use DAX-code in the query editor.

Either use the DAX code in the table view of the data model and add a column there or search for some M-code to use in the query editor.

[Anonymous]

Hi @ImkeF ,

Thank you for the prompt reply, and insight!

So, to make sure I understand, "DAX code" is used within PowerPivot table view, and "M-Code" is used in PowerQuery Editor.  

Is there any difference in processing efficiency between solutions in one or the other, assuming the problem can be solved either way.  Is M-Code generally more restrictive than DAX, capability wise?

Sincerely,

Kristopher

[ImkeF]

Hi @Anonymous ,

yes, you're right.

Performance depends. But most often, DAX is faster.

Disadvantage of a DAX-columns is that it cannot be optimized for compression in the Vertipaq engine. But that effect will most likely only be relevant in very large models.

[Anonymous]

@ImkeF ,

Thank you very much for all of this information, it has been especially helpful and enlightening!!  Apologies again for my inexperience with this method of reporting/coding.

The dataset I am reporting from is relatively large, in comparison to previous dataset's I reported from; it covers several years of ticketing systems data.  Each primary data table contains 300K - 700K rows, appx.  I'm trying to limit the scope of data pulled to only the MVP, to retain efficiency in processing and storage.

Have a great remainder to your week!!

Sincerely,

Kristopher