Previous |
Next |
When you create a dimension object using OWB, the dimension must conform to the following rules:
A dimension must have a surrogate identifier and a business identifier.
The surrogate identifier can consist of only one attribute. However, the business identifier can consist of more than one attribute.
Every dimension level must have at least one attribute.
A dimension attribute can be either a surrogate identifier, a business identifier, a parent identifier, or a regular attribute.
A regular attribute can also play only one of the following roles at a time: effective date, expiration date, or triggering attribute.
A dimension that uses a relational or ROLAP implementation must have at least one level.
Any database table or view that implements a dimension that uses a relational or ROLAP implementation must have only one LONG, LONG RAW, or NCLOB column.
For a dimension that uses a relational or ROLAP implementation, all level attributes must bind to database tables or views only.
A dimension that uses a relational or ROLAP implementation must be associated with a sequence that is used to load the dimension key attribute.
The dimension key attribute of a dimension that uses a relational or ROLAP implementation must bind to the primary key of a table.
A Type 2 Slowing Changing Dimension (SCD) must have the effective date, expiration date, and at least one triggering attribute.
A Type 3 Slowing Changing Dimension (SCD) must have the effective date and at least one triggering attribute.
|
Note: For dimensions with a ROLAP implementation, there are implications and limitations related to the various dimension structures when either reporting on the underlying tables or deploying to the OLAP catalog. Refer to the topic, "Limitations of Deploying Dimensions to the OLAP Catalog" under Overview of Dimensions in Oracle Warehouse Builder Data Modeling, ETL, and Data Quality Guide |
|
See Also: "About the OLAP Catalog" under Overview of Implementing Dimensional Objects in Oracle Warehouse Builder Data Modeling, ETL, and Data Quality Guide |