The Data Vault Handbook - Concepts and Applications

POINT-IN-TIME TABLE EXAMPLE

In this example, we use a PIT table to track patient information at specific moments in time. The PIT table captures data from two satellites connect- ed to the Patient Hub: one storing personal details (such as name, address, and contact info) and the other tracking emergency contact details. These satellites are loaded from two different source systems—the hospital man- agement system for personal details and the emergency services system for emergency contact details. Since updates in each system occur at different times, the PIT table allows us to retrieve a consistent view of the data at spe- cific points in time by indexing these changes.

Take into account that in some instances, upon joining a Hub to one of its Satellites, there can be no corresponding Satellite data available for certain snapshots. This can occur when the business key was either unavailable or unknown to the data source at that given time. To address this issue, “ghost records” are introduced into Satellite entities to fill gaps at the beginning of the timeline, allowing equal joins to function smoothly in queries against the Raw Vault. Ghost records serve as default placeholders, replacing NULL or empty values. For example, the hash keys are typically filled with repetitions