1 Reply Latest reply on Oct 29, 2013 9:38 AM by Matt Lutton

    How do I implement one-to-many master detail relationships properly?

    Brian Nuckels

      We have run into an issue with one-to-many master detail relationships.  The reason we need the child records is to filter our dataset based on these values.  The detail is almost like a "tag" value.  So in our data model, there is a single row of master data with costs, and then each of those rows has 2 child rows which are further "categorizing" the master rows.  We can then use those rows to segment data.  But this varies by customer.

       

      Example

       

      Customer Alpha

      Master ID A cost = 100.00

      Child ID A.1 = Color "Yellow"

      Child ID A.2 = Size "Small"

       

      Master ID B cost = 50.00

      Child ID B.1 = Color "Blue"

      Child ID B.2 = Size "Small"

       

      Customer Beta

      Master ID A cost = 110.00

      Child ID A.1 = Fabric "Cotton"

       

      Master ID B cost = 40.00

      Child ID B.1 = Color "Rayon"

       

      The primary filter in our visualization is Customer.  So you are not seeing customer Alpha and Beta at the same time in visualizations. So within views for Customer A, we can then aggregate costs based on Color or Size.  For Customer B, we can aggregate by Fabric.

       

      Namely, the problem is that this seems to inflate our costs, because the underlying rows in the "master" are multiplied (n * 1) for each child/detail row.  There must be a way to have the visualization properly aggregate but we have not been able to figure this out.

       

      Any ideas?

       

      Dorean Flowers