Star schema
A star schema organizes your data into fact and dimension tables. You can use dimension tables to filter the fact table. Dimension tables contain information that is repeated over and over again, for example, in the Sales table. If you think of a product dimension, it can hold all the information about a product, such as the name, color, SKU, size, and weight. Instead of repeating that information over and over, you can represent the product by an SKU number or even an integer in the Sales table. You can then filter the Sales table by selecting a product name, or even filter it to all products sold that are purple.

Figure 5.1 – An example of a star schema
Complicated models
A complicated model is confusing for report designers and slower for Power BI to refresh and render.

Figure 5.2 – A very complicated model
Notice how hard it is to read and follow the relationships.
Simple model
A simpler model is easier to maintain, create reports on, and refresh.

Figure 5.3 – A simpler model
A simpler table structure should be simple to navigate. The columns and tables should have friendly names, not obtuse system-defined ones. Whenever possible, you should merge or append tables to minimize relationships. Those relationships should also make sense.
Defining relationships
As you can see in both Figure 5.2 and Figure 5.3, there are lines connecting the tables in the modeling view. These lines define the relationship between the tables.
Manage relationships
You can manage relationships by selecting the Manage relationships button on the modeling tab.

Figure 5.4 – Manage relationships window
Using this window, you can create, edit, and delete relationships between tables. You can also choose the Autodetect… button, which will try to find relationships in your data.
Edit relationship
You can choose to edit relationships from Manage relationships or by right-clicking on one of the relationship lines in the model and selecting Properties. You can set or change the properties of any relationship in the Edit relationship dialog window.

Figure 5.5 – Edit relationship dialog
Relationships control filtering and can change the behavior of calculations. They are at the heart of how a Power BI data model works and are one of the key differentiators between Power BI and its competitors. Choosing a subset of data from one side of the relationship can filter what data is available on the other side. If you have a Date table and filter it to today, if that Date table has a relationship with your Sales table, the Sales table will also be filtered. Within the Edit relationship dialog, you can change how the tables interact in two ways: with cardinality or cross-filter direction.