Constraints are required to automatically create entities for data discovery. When you use the auto-create entity wizard, the wizard adds all related child tables for the parent table you select and adds all related tables to the entity. Constraints are also required to search the Data Vault from the entity level. The entity that you search on must have the relationships required to related tables to show related information at the transaction level. Constraints are not needed to archive data. When you run an archive or retirement job, the ILM engine creates a WHERE clause that determines what data needs to be archived.

When you import metadata from a source database to the ILM repository, all constraints that are defined on the database level are imported. If the source database does not include constraint information at the database level, such as in the application level, then you must add the constraints to the ILM repository. You can add constraints manually. Or, you can use tools to help you discover table relationships and create constraints based on the suggested relationships.