faktatabellen

Faktatabellen, in data warehousing and business intelligence, stores the measurable, quantitative data used for analysis. Each row represents a single transaction or event at a defined grain, such as a sale, shipment, or service engagement. In many designs it is the central table in a star or snowflake schema. The term faktatabellen is used in some Scandinavian contexts to refer to this concept.

The faktatabellen is surrounded by dimension tables that describe attributes such as time, product, customer, and geography. The fact table contains foreign keys to these dimension tables and numeric measures called facts, such as sales_amount, quantity_sold, and cost. Facts have additive properties: additive facts can be summed across dimensions; semi-additive facts can be summed along some dimensions but not others (for example, inventory levels over time); non-additive facts cannot be meaningfully summed.

Common types include transaction fact tables (one row per event), periodic snapshot fact tables (aggregates for a fixed period), and factless fact tables (recording events without numeric measures). Design choices such as grain, degenerate dimensions, and the balance between normalization and denormalization affect query performance and ETL complexity.

In practice, faktatabellen enable fast aggregations and drilling across multiple dimensions in BI tools, supporting OLAP-style analysis, reporting, and forecasting. Key design considerations include selecting an appropriate grain, ensuring consistent dimension keys, handling slowly changing dimensions, and maintaining data quality through robust ETL processes. Example structures typically include time_id, product_id, and store_id as foreign keys, alongside measures like sales_amount, units_sold, and discount_amount.