The Data Vault Handbook - Concepts and Applications

of the character “0”, and the timestamp is set to the earliest possible value (0001-01-01 00:00:01).

POINT-IN-TIME TABLE LOGICAL MODEL

When visualizing the logical model, a PIT table is denoted by a ribbon locat- ed over the relevant parent entity. 3.3.2. BRIDGE TABLES Bridge tables in Data Vault are designed to facilitate grain shifts, allowing data to be aggregated or de-aggregated as needed downstream in the data pipeline. These tables are used to adjust the level of detail in queries, making it possible to either increase or decrease granularity for better analysis based on a snapshot date. Unlike Link tables, which typically maintain a specific level of detail, Bridge tables can accommodate a different data grain, even incorporating aggregated values often calculated using GROUP BY statements. These aggregations not only reduce the complexity of downstream queries but also enhance performance, especially when dealing with resource-intensive calculations. Additionally, Bridge tables can also host the business keys of the referenced Hubs if required. By materializing these changes in granularity directly within the Bridge table, the data flow from the Raw Data Vault to the Information Mart becomes more efficient, especially when handling complex calculations or large datasets. Un- like PIT tables, which optimize queries within a single Hub or Link, Bridge tables span across multiple Hubs and Links. They are considered part of the Business Vault because they are system-generated and primarily serve as query optimiz- ers, rather than storing auditable business data.