Forum Discussion

Caitlin_Knox's avatar
Caitlin_Knox
Advocate III
9 years ago
Solved

Data modeling question

I'm hoping someone can at least point me in the right direction to do more research. I have a field in my data set that technically a single line text, but has multiple values separated by commas. How can I extract that information and use it in a meaningful way in my report visualizations. Specifically, I woudl like to use it a slicer since it lists Provider names. Here's a screenshot of what the column looks like

  • Sean's avatar
    Sean
    9 years ago

    Caitlin_Knox  Let me know if you have any questions...

     

    Hope this helps! :smileyhappy:

     

     

12 Replies

  • Phil_Seamark's avatar
    Phil_Seamark
    Microsoft Employee

    Using the wonderful Split Columns option in the Query Editor.  It will create new columns which you can then rename.

     

    • Caitlin_Knox's avatar
      Caitlin_Knox
      Advocate III

      So, that would create 5 columns..Which do I use when creating the slicer? And how would thatrelate back to the rows?

      • Sean's avatar
        Sean
        Community Champion

        Can you post sample data of this column and what your Unique Row Identifier looks like?

         

        Caitlin_KnoxBasically this is how you can do this...

        If something is unclear just ask! :smileyhappy:

  • CahabaData's avatar
    CahabaData
    Memorable Member

    If I assume that you wish to slice on each of those phrases (between commas) - then I'll add to the other replies already posted: your modeling is several steps.

     

    You will first use the split column feature - to set them in their own columns

     

    but then you need to create a new table with your unique record ID and just those columns - and then make that UnPivoted so it is table by itself that has ID in each record and 1 phrase in each record.

     

    This final table needs to be joined back to the original table and is the 'Many' table of a 1:Many relationship.

     

     

    • Caitlin_Knox's avatar
      Caitlin_Knox
      Advocate III

      The unique record identifier is just a number (ID) as its an OData connection to a SharePoint Online list. So I just use the default ID created when a new item is added to the list.

       

      In order to create the new table with the unique record identifier and (now) 5 columns, do I just multi select and then...?

       

      If the answer is just copying and pasting, how does that second table get refreshed when new items are added?

  • If i have a Report connected with One Drive, how i can modeling the data??? I have one fiel as date in the original file but when i use it in Power BI it looks like text how do i can change it?

  • Anonymous's avatar
    Anonymous
    Not applicable

    Hi Sean,

    We have just set up a column in MSProject, within the Projects table, for Systems Impacted. At present only 5 projects have added details of systems impacted, but there are 5 possible systems so the column could have up to 120 variations.

    Does your solution (the only one I've found to do what I need) mean that each time a new project is added or an additional impact is identified, we will need to go through the whole process of splitting/un-pivoting etc ?

     

    Data so far

    Project 1 - Systems affected - Alpha

    Project 2 - Systems affected - Alpha, Beta

    Project 3 - Systems affected - Beta, Gamma, Delta

    Project 4 - Systems affected - Gamma, Delta

    Project 5 - Systems affected - Alpha, Delta

     

    Regards

    Fred

    • Anonymous's avatar
      Anonymous
      Not applicable

      Hi Sean,

       

      As they say, "suck it and see" :smileyhappy:

      I did and the refresh process works perfectly, grabbing new entries each time.  Thanks for the perfect solution.