Explanation:
Box 1: create an active relationship for PaymentDate
In a Microsoft Power BI semantic model with multiple relationships between the same two tables (e.g., FactSales and DimDate), only one relationship can be active at a time.
To select or compare values based on a secondary date, such as PaymentDate, while keeping OrderDate as the primary (active) link, you typically set the PaymentDate relationship as inactive and use the USERELATIONSHIP DAX function within a measure to invoke it.
Key Facts
Active vs. Inactive: Only one active relationship can exist between two tables; additional connections are automatically set as inactive (dotted line).
Invoking Inactive Relationships: To calculate metrics based on PaymentDate (the inactive link), you must wrap your measure in a CALCULATE function:
Total Payments = CALCULATE(SUM(FactSales[Amount]),
USERELATIONSHIP(DimDate[DateKey], FactSales[PaymentDate])).
Filtering Behavior: By default, any filter applied to the DimDate table will only propagate to FactSales through the active relationship (likely OrderDate).
Alternative (Role-Playing Dimensions): Instead of active/inactive relationships, you can create separate "role-playing" date tables (e.g., a DimOrderDate and a DimPaymentDate table) to allow independent filtering in the same visual without special DAX.
Box 2: role-playing
The DimDate table is a role-playing dimension. This is a common data modeling technique where a single physical dimension table is used multiple times within the same fact table, with each instance serving a different logical purpose or "role".
In your Power BI semantic model:
Single Physical Table: You have one physical DimDate table containing all date-related attributes.
Multiple Logical Roles: This single table is linked to the FactSales table through three different foreign keys: OrderDate, PaymentDate, and ShippingDate. Each of these columns represents a different "role" the date dimension plays.
Relationships in Power BI: In Power BI's data model, only one of these relationships can be active at a time (indicated by a solid line in the relationship view), while the others remain inactive (indicated by dotted lines).
Reference:
https://datasturdy.com/active-vs-inactive-relationships-in-power-bi-what-we-need-to-know
https://www.thedataschool.co.uk/svetlana-brazukevich/role-playing-dimensions-in-power-bi