Table Relationships with Different Data Sources

[Anonymous]

Hi, I just recently got introduced to and starting using the OneLake catalog as a data source for my current sales data. I have a fact table that has all my sales for 2025 in the OneLake and then I have other dimension tables that are imported from excel files. Some of those diemsnion tables include product and customer information. 

I am trying to create a sales report that can be sliced/filtered based on product information, customer information, and the sales representative. But I noticed that my 2025 sales table wasn't adhering to these slicers. I have some other fact tables that are imported from excel files and not in the Onelake, and those aren't having any problem being filtered by my slicer visuals. I'm guessing this has to do with a problem in the relationship that is established or with the OneLake data source itself? I'm not quite familiar with the way OneLake works how it effects data modeling but I did notice that the connection for the relationships I created are defined by a different type of line. 

They're almost like parenthesis attached to the end of the relationship lines.

This is what my 2025 sales data is going. All the prices are the same for all the segments, when it should be doing what the Total YTD Sales 2024 column is doing. 

Can anyone tell me what the difference this line means compared to a regular many-to-one relationship? And how I can fix it so my Onelake data source is being filtered/sliced properly?

[lbendlin]

Understanding blank row and limited relationships - SQLBI

Power BI is not sure if your dimension tables cover the fact tables or not. You have to guarantee that yourself.

[Anonymous]

How do I confirm that the dimension tables cover the info in the fact tables? 

[lbendlin]

For the join column make sure that each value in the fact table is present in the dimension table too. Read about "Referential Integrity".

[Anonymous]

I did check and each value in the fact table is present in the dimension tables with the fields I have. All the customers and products I have in my OneLake fact table all are present in my customer and product dimension table. The "Assume Referential Integrity" option in the data modeling is grayed out. 

[lbendlin]

If your check was successful then you can proceed. Keep a watchful eye on the join column cardinality. Ideally it should not exceed 50K

[Anonymous]

The relationship between my fact table and dimension table still shows the parenthesis on the lines, and "assume refeential integrity" is still grayed out. I'm still experiencing the same issues I had before when creating visuals and applying slicers/filters. None of my OneLake data is being sliced/filtered by dimension table fields. 

[Anonymous · Accepted solution]

For those wondering, I ended up using the USERELATIONSHIP function in my report which seemed to work. I think this is only a temporary solution but it allowed for slicing and filtering to occur between those two tables.