Planning Analytics

 View Only
  • 1.  Suppress 0 in Dynamic Reports

    Posted Thu October 14, 2021 12:55 PM
      |   view attached
    Hi,

    I've got a question on suppress 0 using on PAfE Dynamics Reports.

    In a cube, if we have 2 dimension, for instance Version and Year, the zero suppression have a problem :

    Versions : BUD / FCT
    Years : 2020 / 2021

    If we have the values :
    BUD 2021 : 0
    FCT 2020 : 0
    BUD 2020 : 1000

    And build an dynamic reports with just BUD 2021 and FCT 2020, the line is not suppressed with the zero suppression.

    I think this is the normal behaviour of PAfE but colleagues told me that they do not encounter this situation on there applications. 

    In the screenshot in attachment, the cube have just 3 dimension and there is a zero suppression. There is value on EST 2012 REVENUE.

    Corentin

    ------------------------------
    Corentin Gombert
    ------------------------------

    #PlanningAnalyticswithWatson


  • 2.  RE: Suppress 0 in Dynamic Reports

    IBM Champion
    Posted Thu October 14, 2021 01:39 PM
    Hi,

    What you are seeing is expected behaviour.
    The way that row suppression works in Dynamic Reports is by checking whether there is data for the:
    • Row Elements
    • Any Title Element Selectors you have defined (if you have them - your example doesn't)
    • And Column Elements

    The column element is the annoying bit here, because you have nested your columns - it actually takes the distinct lists from each and then essentially unions them:
    • BUD + EST
    • 2012 + 2013
    Even though you are not displaying 2012 + EST together, both elements exist individually in columns - so the system still checks them.


    Good news for the future - IBMers are often talking about the future release of new updated reports that will allow us I believe to Suppress Rows AND Columns. But we don't have that functionality yet.

    In the interim, you can get around it in an example like yours by using an MDX expression on the row dimension instead of zero suppression.
    So you could add MDX to the TM1RptRow:
    {Filter ( 
       { [Measures].members }, 
       [CubeName].([BudDim].[BUD],[YearDim].[2012]) <> 0 
    )} 
    + 
    {Filter ( 
       { [Measures].members }, 
       [CubeName].([BudDim].[BUD],[YearDim].[2013]) <> 0 
    )} 
    + 
    {Filter ( 
       { [Measures].members }, 
       [CubeName].([BudDim].[EST],[YearDim].[2013]) <> 0 
    )} 
    ​


    It can get annoying and complex to try and do it with MDX on bigger reports though.
    You can make the MDX life easier for you in those instances by having Consolidations of the years for each "Version" for example, and then you check whether the intersection of the Year Consolidation combined with the relevant version is not 0. I would not recommend doing individual column checks like I have above if you have more than a few columns!


    Thanks,
    Declan



    ------------------------------
    Declan Rodger
    Technical Director
    Spitfire Analytics
    ------------------------------



  • 3.  RE: Suppress 0 in Dynamic Reports

    Posted Fri October 15, 2021 04:00 AM
    Hi Declan, 

    Thank you, it was the behaviour I expected. 

    I tried the MDX solution but there is a problem : I have 2 dimensions in rows and I'm not able to put the first one in the MDX (The MDX is here to filter the second one). The result is that some elements are missing in the 2nd dimension.

    It's like I construct a MDX in a cube view : if the first dimension is in context filter, MDX consider it as a "CurrentMember", but if the first dimension is in rows MDX do not considere it (or take the consolidation which has 0 value in my case) and do not display elements .. :/ 

    I think I will test with a feeded flag value which takes one if an element in column is not empty and base my Dynamic report on it .. 

    Thank you, 
    Corentin

    ------------------------------
    Corentin Gombert
    ------------------------------



  • 4.  RE: Suppress 0 in Dynamic Reports

    IBM Champion
    Posted Fri October 15, 2021 05:08 AM

    Hi,

    Yes with nested rows and more complex reports it can often be easier to base the report on an entirely different view with a single flag column, hide that column in Excel and then just use a separate view reference to actually have the columns that you want.
    On occasion, in the past I have done this with a completely separate cube, that only serves the purpose of being used to drive which rows are shown in various reports. If your users actually work with the cubes in exploration mode etc. it at least hides the "Display Rows" measure or whatever you may call it, so they don't come asking what it is all the time.

    It does of course become a bit of an overhead to set it all up. If the reports used don't change very often you are ok. If they do, you may have users turning up every week and asking you to create all of these additional rules just to filter some rows!

    Good luck,

    Declan



    ------------------------------
    Declan Rodger
    Technical Director
    Spitfire Analytics
    ------------------------------