I am learning how to use Power BI and one of the repeating pro tips is to avoid bidirectional filtering. I have a model with the following structure:
Table 1 (Course names):
Sample data:
C1 Python
C2 Calculus
C3 Algebra
Table 2 (Course category):
Sample data:
C1 CS
C2 Math
C3 Math
Table 3 (Course lecturers):
Sample data:
L1 C1
L2 C1
L1 C2
L3 C2
L3 C3
There may be multiple courses in a single category and multiple lecturers in a single course.
I would like to generate a matrix report where the rows are by lecturer name, the columns are by category, and the value is the distinctcount of courses. In other words, I would like to understand how many courses from each category each lecturer teaches. For the given sample data I would expect the following result:
CS Math
L1 1 1
L2 1 0
L3 0 2
The only way I got this working was to define the relationship between the tables as bidirectional. Is there an alternative schema that avoids this?
The advice is correct - avoid bi-directional filtering for numerous reasons. Where you need to propagate a relationship, you can turn on bi-directional filtering for a single measure as required using CROSSFILTER().