Data exposure should use the following rules:
- Data Architects and Data Engineers have access to all Domains;
- Product Owners and Business Analysts have access to their own projects data through both the Domain project (raw data) and the Product project (business oriented objects);
- All teams/users outside the project that are not either Data Architects or Data Engineers will ALWAYS access third party domains data (either RAW or business oriented) through their own Product projects. This means that the cost associated to querying third party domains will always be associated to the query owner, not to the data owner.
In terms of exposure method:
- Each Domain must have at least one Product project
- For each Dataset created in the Product project, an "twin" Dataset must be created in the Domain project
- Information can exposed in one of the following ways:
- A table is created directly in the Product project Dataset, and filled in there
- A view is created in the "twin" Domain Dataset, pointing to the Domain DIM and FACT Table. A second view is created in the Product project Dataset, pointing to the "twin" Domain Dataset view. The product service account is granted access as reader to both the "twin dataset" in the domain project, as well as to the product project dataset.
