The Lost Art of Building Bridges

As a data vault grows with your business whilst by solving multiple business requirements it may be time to consider the need to build a bridge between the business objects to serve an information mart requirement. A bridge table in data vault is one of the two constructs designed to serve an information mart layer of your analytics platform by

• Using keys to take advantage of the platform’s features to improve querying performance
• Taking the complexity of performing these joins repetitively away from the business users
• Shortening the distance between hub tables

The other query assistance construct is the point-in-time (PIT) table which we have discussed quite a bit before. Like the PIT table the bridge table is disposable, which is why it is not a data vault table, but rather a construct designed to leverage both raw and business vault to optimize query performance from data vault and to simplify SQL joins for the general information mart user. We also have the option to include lightweight derivations/aggregations/calculations in the bridge table, nothing too complex. Basically, the difference between including calculations in a bridge table versus using a business vault link table is threefold:

• Is the business rule outcome idempotent?
• Do we need to maintain…