Forum Discussion

luxpbi's avatar
luxpbi
Icon for Helper V rankHelper V
8 years ago

Power BI Desktop Extremely Slow

Hi, The fact is that I have a not very big pbix. It's 32MB size. I've create a calculated column, a little bit complex, but it works. The problem is that since then the pbix is extremely slow, for...
Need Help
Kumail's avatar
Kumail
Icon for Impactful Individual rankImpactful Individual
8 years ago

Hi luxpbi

 

I have tested the dax expression. Initially, it was giving number of syntax errors, however, it worked with some minor tweaks. The image and file is attached for your reference.

 

Regards

Kumail

Analytics and BI Practitioner

Email: [email protected] Skype: kumailrazakazi

Kumail's avatar
Kumail
Icon for Impactful Individual rankImpactful Individual
8 years ago

Just figured that file upload option is not available here. 

May be you can contact on my email id for file or any other option to share you the file.

 

Regards

  • luxpbi's avatar
    luxpbi
    Icon for Helper V rankHelper V
    8 years ago

    Hi Kumail,

     

    Thank you for your time and your answer, but I can't really se the code there, maybe you could post it in a comment.
    In the meantime I have made some modifications and I made it simplier and it works fine.

    Days =
    VAR CurrentProduct = 'Table1'[Product]
    VAR CurrentDate = 'Table1'[Date]
    VAR CurrentNMov = 'Table1'[Index_]
    VAR NextStateDate =
     CALCULATE (
     MIN ( 'Table1'[Date] );
     FILTER (
     'Table1';
     'Table1'[Product] = CurrentProduct
     && 'Table1'[Date] >= CurrentDate
     && 'Table1'[Index_] > CurrentNMov
     )
     )
    RETURN
     IF (
     ISBLANK ( NextStateDate );
     ( TODAY () - CurrentDate )
     + 1;
     ( NextStateDate - CurrentDate )
     + 1
     )


    What I'm thinking is that maybe instead of a calculated column I could use a measure.
    Have you tested my original code and yours with DAX Studio?

    Regards,

  • Kumail's avatar
    Kumail
    Icon for Impactful Individual rankImpactful Individual
    8 years ago
    TestDays = 
VAR CurrentProduct = 'Table1'[Product]
VAR CurrentState = 'Table1'[Status]
VAR CurrentDate = 'Table1'[Date]
VAR CurrentNMov = 'Table1'[Index]
VAR CurrentCodTaller = 'Table1'[Type]
VAR NextStateDate =
    CALCULATE (
        MIN ( 'Table1'[Date] ),
        KEEPFILTERS ( 'Table1'[Status] <> 8
            || 'Table1'[Status] <> 9 ),
        FILTER (
            'Table1',
            'Table1'[Product] = CurrentProduct
                && 'Table1'[Date] >= CurrentDate
                && 'Table1'[Index] > CurrentNMov
                && ( 'Table1'[Type] = BLANK () || 'Table1'[Type] <> CurrentCodTaller || 'Table1'[Type] = CurrentCodTaller )
        )
    )
VAR LastMov =
    CALCULATE (
        MIN ( 'Table1'[Date] ),
        FILTER (
            'Table1',
            'Table1'[Product] = CurrentProduct
                && 'Table1'[Date] >= CurrentDate
                && 'Table1'[Index] > CurrentNMov
        )
    )
RETURN
    IF (
        ISBLANK ( LastMov ),
        ( TODAY () - CurrentDate )
            + 1,
        ( NextStateDate - CurrentDate )
            + 1
)

    Even in measure, its working perfectly fine; as fast as any other feature in power bi

    • luxpbi's avatar
      luxpbi
      Icon for Helper V rankHelper V
      8 years ago

      Works fast because of the size of the dataset. When you have 2 million rows is veeeery slow.. 
      Do you think is there a way to optimize it?

      • Kumail's avatar
        Kumail
        Icon for Impactful Individual rankImpactful Individual
        8 years ago

        Not sure. Actually I was extremely busy on something therefore, couldn't reply you yesterday.

         

        I have used python for my clients for processing large dataset and created a csv or excel file for visualization (If you are not querying data directly from CRM). That generally works well for me.