Overview of the OLAP Catalog

The repository for OLAP metadata is known as the OLAP Catalog. OLAP metadata represents warehouse data as logical cubes, as described in "OLAP Metadata Model".

OLAP metadata must be defined and mapped to any data that will be accessed by the OLAP API. OLAP metadata may also be used by other types of analytical applications.

OLAP metadata maps to dimension tables and fact tables. The dimension tables must be organized in levels. The dimension tables and fact tables may be actual relational tables or they may be views representing data stored in analytic workspaces. A number of different warehouse configurations can be represented by OLAP metadata, as described in "Data Warehouse Requirements".

Tools for Creating OLAP Metadata

There are two tools for creating OLAP metadata:

• Oracle Enterprise Manager.
• The CWM2 PL/SQL APIs.

  Note: Enterprise Manager currently uses a set of proprietary APIs to create OLAP metadata. It does not provide access to metadata created with the CWM2 APIs. However, the OLAP Catalog metadata views allow you to browse all the metadata in the OLAP Catalog. This includes metadata created by Enterprise Manager and metadata created by the CWM2 APIs.

OLAP Catalog Components

The OLAP Catalog includes the following:

• Metadata model tables - A set of tables that instantiate the OLAP metadata model. These tables define all the OLAP metadata objects: dimensions, measures, cubes, measure folders, and so on. Within the metadata definitions are references to the actual warehouse data.
• A Write API - A set of PL/SQL packages for creating and editing OLAP metadata. These packages contain procedures for inserting, updating, and deleting rows in the model tables.
• A Read API - A set of SQL views providing information about the metadata registered in the model tables.

Logical Steps for Creating OLAP Metadata

Whether you create OLAP metadata programmatically or by using Oracle Enterprise Manager, you follow the same logical steps.

To create OLAP metadata:

1. Create the dimensions. Specify the levels, attributes, and hierarchies associated with each one.
2. Create cubes and specify their edges (dimensions).
3. Create measures that represent the fact data. Associate each measure with a cube.
4. Map the metadata entities to the source data.
5. Create measure folders in which to store related measures. Populate the folders with measures.