Forum Discussion

emarome94's avatar
emarome94
Icon for Helper I rankHelper I
3 years ago
Solved

Calculate Statement - Slow Table Load with many informations

Hi guys,

I have created this measure (Measure1) that actually obtain the result I want to discover and gives not many problem in limited filter context like table with few rows. 
Measure1=
CALCULATE(

    [Measure0],
    FILTER( 'Accounting', 'Accounting'[Fiscal Period] >= RELATED( 'Price Lists'[Fiscal Period] )
))
The problem comes when I add more information in the table report like customer, item, country etc. lead to a very slow loading. 
The expression within calculate is made by the following measure:
Measure0 =
SUMX ( 
    FILTER ( 'Price Lists', 'Price Lists'[Old Price]> 0 ),
IF (
        [ASP] > [Old Price],
        ( [ASP] - [Old Price] ) * [Quantity],
        BLANK() ))

From the Measure 0 I am creating a context where from the price list I am creating a table with only the values "old price" area greater than zero and then creating and then the if statement.
The measure 1 is using the measure 0 only where is retrieving the fiscal period from the price list with related. 

 

Can someone help me maybe with a best practice to fix the formula?
Really thank you in advance for your help,

Regards,

ES

CALCULATE
slow rendering
Slow Reports

5 Replies

  • ValtteriN's avatar
    ValtteriN
    Icon for Community Champion rankCommunity Champion
    3 years ago

    Hi,

    The issue is likely caused by your use of RELATED and additionally the use of complex logic within IF statement. RELATED function often suffers from performance issues and it should be avoided when possible. In your dax you use related to filter upcoming fiscal periods for calculation:    FILTER( 'Accounting', 'Accounting'[Fiscal Period] >= RELATED( 'Price Lists'[Fiscal Period] )

    The first question to ask here is if you can achieve the same calculation without using RELATED e.g. using MAX:

    Measure1=
    CALCULATE(

        [Measure0],
        FILTER( 'Accounting', 'Accounting'[Fiscal Period] >= MAX( 'Price Lists'[Fiscal Period] )
    ))


    One other thing to consider is if you need FILTER here. often you can achive the same results just by dropping it. 

    e.g. 



    Measure1=
    var _period = MAX( 'Price Lists'[Fiscal Period] )
    CALCULATE(

        [Measure0],
       'Accounting'[Fiscal Period] >= _period)

    Generally speaking avoid using RELATED and FILTER and your calculations should be faster.

    I hope this post helps to solve your issue and if it does consider accepting it as a solution and giving the post a thumbs up!

    My LinkedIn: https://www.linkedin.com/in/n%C3%A4ttiahov-00001/

     

    • emarome94's avatar
      emarome94
      Icon for Helper I rankHelper I
      3 years ago

      Hi ValtteriN,

       

      Thank you for your quick response.

       

      Unfortunately, I can not reach the same result with the MAX measure (in this case it would retrieve the higher fiscal period in the price list fact table) where instead I need the exact fiscal period for start the calculations. 

       

      For instance if in the price list table there is the customer "BigCompany" with item "RedItem" with fiscal period 2023004, I need to start my calculation from FY 2023004 ( >= related (price list[fiscalperiod]) )

       

       

      • ValtteriN's avatar
        ValtteriN
        Icon for Community Champion rankCommunity Champion
        3 years ago

        I see, maybe you could calculate the desired fiscal year using CALCULATE()? It is hard to provide example fitting your use case without seeign the data, but the basic idea is as follows:

        Instead of using RELATED you would use the same filter context in variable
        var _FY = 
        CALCULATE(MAX( 'accounting period'[Fiscal Period]), ALL(accounting period[Fiscal Period]), 'accounting period'[Fiscal Period] =MAX( 'Price Lists'[Fiscal Period]))
        return 

        CALCULATE(

            [Measure0],
           'Accounting'[Fiscal Period] >= _FY)



        This dax might or might not work depending on your data, but the basic idea should function with some changes.