In 10g, a difference was made between:
From 11g, no difference in the name is made and this two appellations are regrouped under subject area.
Even though the Logical SQL requests from Answers and other clients query the presentation tables and columns, the real logic for entities, relationships, joins, and so on is in the Business Model and Mapping layer.
The primary function of the Presentation layer is to provide custom names, dictionary entries, organization, and security for different groups of users.
A subject area contains:
- presentation tables,
By default, it's created from one logical fact table and it's related dimension.
Although each subject area must be populated with contents from a single business model, you can create multiple subject areas for one business model. For very large business models, you may want to do this to help users work with the content.
One important reason to use a custom Presentation layer is to make the schema as easy to use and understand as possible. Therefore, users should not be able to view columns that have no meaning to them. The following columns are examples of columns that you might want to remove from the Presentation layer:
- Key columns that have no business meaning
- Columns that users do not need to view (for example, codes, when text descriptions exist)
- Columns that users are not authorized to see
Only for 11gUsers in Oracle BI Answers can create queries that span multiple subject areas, as long as the subject areas correspond to the same business model.
In the BMM from the version 11g, you can now create automatically the subject area.
In a data warehousing environment, a subject area is a classification system that represents or distinguishes parts of an organization or areas of knowledge. A data mart is often developed to support a subject area such as sales, marketing, or geography.
A consideration to take when designing a subject area is to pay attention at the final user.
Do you design for a user:
- that will use answer to perform ad hoc queries
- or who will interact with the metadata via logical sql, dashboard and prompts.
- Separate each level of a hierarchy in one presentation table
- No more than 11 columns in a presentation table
- Always two identifiers in the name of a column
- Measures by type (Aggregation, Level Based, Time series)
- Only columns that are used logically in the reports
- Only one fact table (one combination of dimension values ie domain)
- Set the Implicit Fact Column (to enable constraints in a dashboard prompt and may be limited choice in the filter wizard).