Slicers based on dimensions filtering other dimensions
I have two excel files (my two fact tables) as a source with common columns.
I've append these two fact tables and created dimensions based on these appends, selecting the columns that I wanted and removing the duplicates and then connecting to my two fact tables.
If I filter the information in the slicers, it filters correctly the two fact tables but it is not filtering between the dimensions (which I understand because there are not relationship between the dimensions and I cannot cross filter)
So how is to make it possible filter the two fact tables and and also the dimensions filtering each other? What I need to change in my data model?
Changing the direction of the relationship between tables from single to both should work.If it doesn't meet your requirement, kindly share your sample data and excepted result to me. Please upload your files to One Drive and share the link here.
Community Support Team _ Frank If this post helps, then please consider Accept it as the solution to help the others find it more quickly.
Workaround found in the official Power BI documentation "Bi-directional relationship guidance" for the data modeling, Chapter "Slicer items with data" . Without changing anything to the model (no bi-directional filter needed, therefore no risk of ambiguity in the model), but with the right filter setup in your (slicer) visual, I was able to solve a similar issue.
There's a better way to achieve the same result: Instead of using bi-directional filters, you can apply a visual-level filter to theProductslicer itself.
Let's now consider that the relationship between theProductandSalestable no longer filters in both directions. And, the following measure definition has been added to theSalestable.
Total Quantity = SUM(Sales[Quantity])
To show theProductslicer items "with data", it simply needs to be filtered by theTotal Quantitymeasure using the "is not blank" condition.
Of course, you don't solve so the issue on the modeling level, but on the reporting layer. I found this solution anyway straightforward for my case.