Snowflake schema vs Many to many relationship

[Lorili]

I have created a new dimension in my model, which is based on a snowflake schema (extended star schema).

My new dimension is called kredietlimiet, and is a "satellite" dimension of the Verkooporder dimension.

So the relationships are as follows: Fact (many to one) Verkooporder (one to many) Kredietlimiet, with bidirectional crossfiltering on.

I noticed that I could not create a visual with a field from the Kredietlimiet dimension and from another dimension Klant in my example). 

I could only make this work by adding a measure. Not necessarily a measure from my fact table, but just a measure:

This seemed strange to me, so I decided to make a direct relationship between my Fact and Kredietlimiet table (the key is the verkooporder ID, the same as in the above example). This relationship is a many to many relationship.

 Then when I tried to do the same thing, it worked without the measure: 

Why is there a difference between these same approaches? I would assume that the relationships would implicitly work the same way. 

 

[Anonymous]

Hi @Lorili 

I tried to restore the scenario you mentioned but failed, can you provide your pbix file(remove sensitive info) or create a sample with simple data for reference ?  

Best Regards,
Community Support Team _ Ailsa Tao