13 Data Vault Terms you Need to Know

Whether you are an experienced data professional or completely new to the industry, understanding Data Vault and its core concepts could be really helpful to your organization.

In this article, we will explore the core principles of Data Vault, exploring 13 terms that every data professional needs to know.

1 – Data Vault

A method & architecture for enterprise data analytics, supporting business intelligence, data warehousing, analytics & data science. Based on the best parts of Kimball and Inmon data warehousing methods, Data Vault adds many improvements. It solidifies best practices in the industry to provide a scalable, effective architecture for integrating many systems with large quantities of data.

2 – Hub

A core component of Data Vault, this table structure stores unique business keys (identifiers) for a business concept (e.g. Customer IDs). Think of Hubs as the starting point for building relationships within the data warehouse.

3 – Link

Associates one or more business concepts (Hubs) by storing the sets of related business keys for those concepts. Together with Hubs, Links form the spine of the Data Model which is used to represent the core business concepts and how they relate and interact with each-other.

4 – Satellite

‘Attached’ to Hubs or Links, Satellites contain the descriptive data about the concept represented by the parent Hub or Link. They track changes in this descriptive data and store the history of those changes over time.

5 – Point-in-Time (PIT)

PITs is optional ‘query helper’ structures which sit between the business vault and the presentation layer. They provide performance improvements for frequently queried sets of Satellites and make it easier to access historical snapshots for a given time for temporal analysis & trend identification.

6 – Bridge Table

Bridges are another optional ‘query helper’ structure used to traverse the spine of the data model (Hubs and Links) to provide re-usable queries around specific relationships and concepts that may be frequently queried or used downstream.

7 – Business Vault

The layer between the raw vault and the presentation layer, this is where business rules (calculations and other transformations) are applied. This layer is the home of most of the business logic in a Data Vault solution, facilitating actionable insights.

8 – Raw Vault

The layer after the staging layer where data is integrated using Data Vault modeling. It serves as the foundation for downstream business logic and querying by providing a business-oriented view of the data, modelled on concepts and processes in the business.

9 – Business Key

A unique identifier used to represent a business concept or entity within the business. It’s the primary reference point for data integration and analysis.

10 – Surrogate Key

System-generated identifiers used to uniquely identify records.

11 – Hash Key

A hashed representation of a business key used for query optimisation and consistency. Usually MD5, though SHA-1 is the recommended standard to avoid chances of hash collisions at high volumes.

12 – Hash-diff

A single column formed by concatenating many columns (usually the Satellite Payload) and hashing them. This acts as a checksum – if one column’s value changes then we know a Satellite record has changed. This is the main way we track history in Satellites and saves us having to compare columns one at a time.

13 – Data Vault Automation

Tools and processes to automate creation, maintenance, and management of Data Vault structures. We even built our own – AutomateDV on the dbt platform!

Conclusion

In conclusion, mastering these essential concepts of Data Vault empowers data professionals to build a robust, scalable, and future-proof Data Vault. By embracing Data Vault, you can unlock the full potential of your data assets, driving informed decision-making and sustainable growth in today’s data-driven landscape.

Need help implementing Data Vault? We offer comprehensive support at every stage of your journey. Learn more about our services here.