faktataulu

A faktataulu, or fact table, is a central component of dimensional data warehouse schemas. It stores numerical measurements or metrics about business processes, such as sales amounts, quantities, costs, or profits. Each row represents a measurable event or observation at a defined grain, and the table includes foreign keys to related dimension tables that provide context (such as date, product, store, or customer).

Structure and grain define the fact table. The key columns are the fact measures (the numeric values) and foreign keys to dimension tables. The grain determines the lowest level of detail captured in the table—for example, one row per sales transaction, or one row per product per day. Dimension keys are usually surrogate keys. Degenerate dimensions, such as an orderID, can appear as a column if relevant. Fact tables may also include additive, semi-additive, or non-additive measures, influencing how calculations across time or aggregates are performed.

Types and schemas. Fact tables are commonly used in star schemas, where they reference a set of denormalized dimension tables, or in snowflake schemas with expanded dimensions. Some implementations include factless fact tables, which record events that have no numeric measures but still link dimension keys to track occurrences.

Example. A sales fact table might include date_key, product_key, store_key, customer_key, and measures such as sales_amount, units_sold, and discount_amount. This enables queries to analyze performance across time, products, stores, and customers.

Maintenance and use. Fact tables are typically append-only in data warehouses, with data loaded through ETL processes. They are optimized for read-heavy queries and summarized by aggregations, partitions, and indexing to support business reporting and analytics.