Which of the following best describes a fact table?

• A. Descriptive attributes stored to constrain or categorize facts.
• B. Stores measurable numeric metrics and keys to dimensions, with grain context.
• C. Normalizes all dimensions into many small tables.
• D. Records only events without numeric measures.

Answer: (B)

A fact table is the central place in a data warehouse that holds the numeric measurements you want to analyze, along with the keys that link to your descriptive dimensions. Each row represents a specific instance at a defined level of detail (the grain), such as one product sold on a particular day at a specific store, and it stores the measurable metrics like units_sold, revenue, or cost. The dimension keys (e.g., date_id, product_id, store_id) connect this row to the descriptive attributes stored in dimension tables, enabling rich analysis while keeping the facts at a precise, consistent level of detail.

This is why the option describing a fact table as storing measurable numeric metrics and keys to dimensions, with grain context is the best fit. Descriptive attributes intended to constrain or categorize facts belong to dimension tables, not the fact table. Normalizing all dimensions into many small tables is a design choice for dimensions, not a defining feature of a fact table. Recording events without numeric measures describes a factless fact table, which is a special case, not the standard description of a typical fact table.