Model Data & Relationships

In a star schema, which statement best describes a dimension table?

• A It stores the numeric measures and foreign keys at the grain of each event
• B It stores descriptive attributes used to filter and group facts, with a unique key ✓
• C It is always larger than the fact table
• D It must use a composite key of all its columns

Answer: Dimension tables hold descriptive attributes (e.g. Product, Date, Customer) with a unique key used to slice and group the numeric values stored in fact tables.

Two tables are related on a key where the dimension side has unique values and the fact side has many repeats. What cardinality does Power BI assign?

• A Many-to-many
• B One-to-one
• C One-to-many ✓
• D Many-to-one only when sorted

Answer: A unique key on the dimension and repeated values on the fact produces a one-to-many (1:*) relationship, the standard pattern in a star schema.

By default, in what direction does filtering flow across a one-to-many relationship in Power BI?

• A From the many side to the one side
• B From the one side to the many side ✓
• C In both directions automatically
• D No filtering occurs until cross-filter is enabled

Answer: The default single cross-filter direction flows from the 'one' (dimension) side to the 'many' (fact) side, so dimension selections filter the related facts.

A model has two relationships between Sales and Date (OrderDate and ShipDate). Only one can be active. How do you use the inactive relationship in a single measure?

• A RELATED()
• B CROSSFILTER() set to None
• C USERELATIONSHIP() inside CALCULATE ✓
• D TREATAS() on the Date table

Answer: USERELATIONSHIP, used as a CALCULATE modifier, temporarily activates an inactive relationship for that calculation, letting one measure use ShipDate instead of the active OrderDate.

Why is a single dedicated Date dimension marked with 'Mark as Date Table' recommended over relying on auto date/time?

• A It forces DirectQuery mode for dates
• B It automatically translates date formats per user locale
• C It provides a continuous calendar and reliable keys so DAX time-intelligence works correctly across all fact tables ✓
• D It disables all relationships to reduce model size

Answer: A marked date table gives a complete, contiguous range of dates and a recognised date key, which DAX time-intelligence functions require, and it avoids the hidden per-column auto date/time tables that bloat the model.

Setting a relationship's cross-filter direction to 'Both' (bidirectional) carries which main risk?

• A It can create ambiguous filter paths and unexpected results, especially with multiple related tables ✓
• B It permanently deletes the inactive relationship
• C It disables row-level security entirely
• D It always doubles the refresh time

Answer: Bidirectional filtering can introduce ambiguity and circular filter paths in models with several relationships, producing unexpected aggregations; it should be used sparingly and intentionally.