Cognos Analytics

Cognos Analytics

Connect, learn, and share with thousands of IBM Cognos Analytics users! 

 View Only
Expand all | Collapse all

Dynamic cube designer - Unable to establish outer join between fact and Dimension table.

  • 1.  Dynamic cube designer - Unable to establish outer join between fact and Dimension table.

    Posted Mon August 31, 2020 12:02 PM
    Hi,

    The funder is not a mandatory column in the KPI Fact table. The requirements for this project are to include all KPI values therefore an outer join relationship has to be established between the Dim_Funders(0) and KPI_Fact(n) table.

    When I try to establish an outer join, Dynamic cube designer highlights it as an issue with the error message.

    "One or more attributes that are used as level keys or the caption in the level Province reference column(s)
    in table DIM_FUNDERS. The outer join allows empty rows to be returned from it when the dimension is populated during cube startup which will result in null members being created."

    The levels defined for funders are

    Province
    Category
    Funders.

    The "Outer Join" relationship is available in the Framework manager & Cognos Transformer.
    Can the "Outer Join" relationship be established in Dynamic Cube Designer? 

    Please advise.

    ------------------------------
    Vincent Dsouza
    ------------------------------

    #CognosAnalyticswithWatson


  • 2.  RE: Dynamic cube designer - Unable to establish outer join between fact and Dimension table.

    Posted Thu April 25, 2024 03:39 PM

    Did you ever get an answer to this question?



    ------------------------------
    Matt Thiese
    ------------------------------



  • 3.  RE: Dynamic cube designer - Unable to establish outer join between fact and Dimension table.

    Posted Mon April 29, 2024 03:26 AM

    Hi, 

    Dynamic cubes needs inner joins. But that's not a problem, the way to handle it, is  when you create you dimension, you would always create a line for the missing values. 

    Make a row in your dimension, let the key be "-1", and the "Funder" value could be "Blank".

    In the fact table you simply replace the lines with missing  "Funder" values with "-1" when loading it in the datawarehouse.

    This way you never loose data, and you can use inner joines, an better boost the performance.



    ------------------------------
    René Kent Nielsen
    Brand manager
    CogniTech A/S
    Herning
    ------------------------------