Multidimensional Diagrams

A multidimensional data diagram provides a graphical view of your datamart or data warehouse database, and helps you identify the facts and dimensions that will be used to build its cubes.

Numeric values or measures such as sales total, budget, and cost, are the facts of the business, while the areas covered by the business, in terms of geography, time, people and products, are the dimensions of the business. A multidimensional diagram shows the facts, surrounded by their dimensions, which will be used to populate cubes for enterprise information management, query and analysis tool and enterprise reporting. In the following example, the Sales fact is surrounded by the Product, Time, Customer, and Store dimensions to allow sales data to be analyzed by any of these criteria:

PowerDesigner maps facts and dimensions to their original operational database tables to enable population of the cubes (see Operational to Warehouse Data Mappings).

• Multidimensional Diagram Objects
  PowerDesigner supports all the objects necessary to build multidimensional diagrams.
• Identifying Fact and Dimension Tables
  When designing a data warehouse, you will need to identify which of your tables and views represent facts (containing numerical values such as sales, revenue, or budget figures), and which dimensions (providing ways of aggregating these figures, such as by region, date, customer, or product). PowerDesigner can retrieve the multidimensional type of a table by analyzing the references attached to it, where child tables or views are identified as candidate facts and parent tables or views are identified as candidate dimensions.
• Generating Cubes
  PowerDesigner can generate facts and dimensions from your operational tables to create a multidimensional diagram representing a cube. The generation will create mappings between your operational and warehouse objects as the basis for extraction scripts or in preparation for generating a BusinessObjects universe.
• Facts (PDM)
  Facts define the focus of the data to be analyzed and how it is calculated. Examples of facts are sales, costs, employee hours, revenue, budget. Facts contain a list of measures, which represent the actual numerical data, and are surrounded by dimensions, which control how that data will be analyzed.
• Dimensions (PDM)
  A dimension is an axis of analysis in a multidimensional structure. Typical dimensions for a sales database include time, region, department, and product.
• Associations (PDM)
  An association connects a fact to the dimension that defines it.
• Operational to Warehouse Data Mappings
  Data warehousing requires the extraction, transformation, and loading of data from operational systems to a data warehouse database. You can create mappings between operational and data warehouse data and from the data warehouse data and OLAP cubes. To review or edit these mappings, open your multidimensional diagram, and then select Tools > Mapping Editor.
• Generating Data Warehouse Extraction Scripts
  You can model operational and data warehouse data structures in PDMs, and specify mappings between the operational data sources and the data warehouse to generate extraction scripts to populate the data warehouse with operational data.
• Generating Cube Data
  You can map physical tables (including those of type dimension or fact) to cube dimensions or cube measures in OLAP databases, and use these mappings to generate cube data in text files to be loaded by OLAP engines. When you use the Rebuild Cubes command to create cubes and dimensions from fact and dimension tables, mappings between source tables and OLAP objects are automatically created.