1 Reply Latest reply on Nov 8, 2015 6:05 AM by Alexander Mou

    The Importance of Granularity (…to Blend or not to Blend?)

    Simon Runc

      In the last couple of weeks I've had quite a few 'Blending' related questions. In many cases the problem has not been 'technical' so much as Blending was being used a substitute for 'Joining'. As always with Tableau, we can work round these

      LOD from multiple Sources?

      but as you'll see, things become increasingly (and unnecessarily) complicated. By choosing the 'correct' option (either Blend or Join) upfront, it makes life so much easier (and more fun!).


      To explain how I see things I wrote the following Blog, on when to 'Blend' and when to 'Join'


      The Importance of Granularity (…to Blend or not to Blend?) | The Data Animators


      In this article I argue that I see the requirement for Blending, when Data resides in different data-sources, as a 'fringe' case...


      ....it’s nearly always possible (by hook, or by crook!) to get data into a single schema (even if the tables need to be materialized into text files and those files placed in the same file location), and do a proper join (if appropriate)


      In light of the revelation from TC15, that in the 'near' (define that how you will!) future we will be able to join across different database schema, this case will be fully redundant.


      In the article, I run through a 'typical' example of 2 data-sets I want to analyse, and show what happens when we 'Join' them, and when we 'Blend' them.


      ...as with most data problems, the answer to the question...


      Should I Join, or should I Blend?


      lies in understanding the 'Granularity' of the data.


      Hope you find this useful, and after reading this, like me, you come to love the way Tableau 'Blends' and are glad it doesn't act like a 'Join'.