Dealing with circular references Similar to how synthetic keys are created, a circular reference can also be the result of unwanted associations in our data model and, as such, they can be fixed using the same principles described earlier. But before getting into how to solve them, let's first see what they are. We can think of the data model as a map that shows the paths through which we can walk to get from one point to another. For instance, considering the data model seen in the previous section, the only path to get from the fact table to the Calendar table is the YearMonth route. This means in order to get there you must know what Month and Year you are looking for in the Calendar table. However, when the data model becomes more and more complex, with a larger number of tables, and more and more destination points, we might also get to a point where we have more than one route connecting point A to point B. All roads lead to Rome, they say. Well, in our case, we must always have one road between any two points. Otherwise, we would be having a circular reference.