Tableau provides users with three different ways to combine datasets, but it can be confusing to know which one to use. In this blog, I will explain the differences between Join, Relationship, and Data Blending in simple terms.

But before we dive into the methods, let’s talk about Tableau’s data model. It has two layers: logical and physical.

The logical layer is where you combine tables using relationships.

The physical layer is where you can use joins and unions to combine data.

Here is another way of visualizing physical and logical layer of Tableau:

Now, let’s get into the methods:

  • A relationship links tables without actually combining them into a single table. Unlike joins, relationships don’t change the underlying structure of the tables, hence maintaining data integrity and avoiding duplication.
  • Join creates a new table that contains all the columns from both tables, but only the rows that match the specified join criteria. This can cause data loss or duplication if tables are at different levels of detail, and joins must be established before analysis can begin.
  • Blend does not combine the data directly. Instead, blends query each data source independently, aggregate the results to the appropriate level, then present the results together visually in the view. Because of this, blends can handle different levels of detail and also work with published data sources. However, blend only supports left join, which means the primary table should contain all possible values.

So, depending on your data, you might want to use a relationship, a join, or data blending. Understanding the differences between these methods will help you choose the right one for your analysis. I hope this helps!

The Data School
Author: The Data School