Dimensions are one of the two main object types in dimensional modeling, alongside fact tables. A Star Schema consists of fact tables, which describe business processes like sales transactions or stock levels, and dimensions, which contain descriptive data such as dates, customer information, or product details.

There are three main types of dimensions: Regular Dimensions, which are the most common and contain information to describe business entities like customers or products; Junk Dimensions, which combine multiple small entities into one dimension to avoid confusion; and Degenerate Dimensions, which are pieces of information stored directly in the fact table, such as transaction numbers.

Regular Dimensions follow specific rules: each row explains one occurrence of an entity, is uniquely identifiable, and contains denormalized data with hierarchies and additional attributes. Junk Dimensions group together business entities with few rows, while Degenerate Dimensions relate directly to transactions in the fact table and are typically not stored in separate dimension tables.

Conformed Dimensions are used by multiple fact tables, such as a product dimension used by both online and retail sales facts. When planning a dimensional model, a Dimension Matrix or Bus Matrix helps connect fact tables to dimension tables, defining the data model structure and reporting axes.