3 Replies Latest reply on Oct 18, 2013 6:51 AM by John Sobczak

    Blending Considerations vs Table Join on 2 Related Data Sources with One to Many Relationship

    John Sobczak

      I use Tableau blending all the time for separate data sources with many to many relationships.  There are also some compelling use cases for using blending for very targeted situations that may have one to many relationships such as a method for computing distinct counts.

       

      But what about consideration of blending 2 data sources that are related by a one to many relationship that would normally be joined?

       

      For example I have 2 data tables - one of insurance policy contracts and a 2nd table on claims on those insurance policies, thus there being a one to many relationship between the 2 tables by policy #  or in other words, possible multiple claims per policy.  And also the claim table generally has 10% the number of rows as policy table.  This sounds like it would be a classic text book table join exercise (including all policies and all claims) resulting in a larger combined table, albeit with a large population of nulls.

       

      However what are the benefits (if any) as well as the negatives, performance or functional, of keeping them separate and blending them through Policy #? I realize that 2 separate independent queries are being performed against each data source and those results combined.