You can build a cube with the Cube wizard or the Cube editor. The Cube wizard takes you through the process. The Cube editor allows you to perform some of the steps in your own order.
To build a cube with either the Cube wizard or the Cube editor, you must specify your cube’s:
Select the data source that contains the fact table and dimension tables you want in your cube. A data source name identifies a database resource and parameters for its usage.
Select the fact table that contains the measures to be included in your cube. A fact table is the central table in a schema. It contains the numerical data (that is, measures) of main interest to users of the cube.
A fact table contains foreign keys that are joined to primary keys in dimension tables.
Select the measures to be available to users of your cube. A measure contains numerical data (for example, Sales) viewed and analyzed by users. Each measure corresponds to a column in the fact table. This column supplies the measure’s values.
Select the dimensions to be available to users of your cube. Dimensions are descriptive categories by which the measures can be separated for analysis. In tabular browsers, they provide the column headings, row headings, and subheadings by which the measures are separated and displayed to cube users. (In graphical browsers, they provide other types of descriptive labels but with the same function as in tabular browsers.) For example, the measure is Sales, and the dimensions are Time, Location, and Product. Users can separate Sales into various categories of Time, Location, and Product. Time provides headings for individual years and subheadings for months. Location and Product also supply a variety of headings and subheadings.
Each dimension is created from one or more columns in a dimension table. These columns supply the dimension’s values which produce the column headings, row headings, and subheadings seen by cube users.
Each dimension table contains a primary key that is joined to a foreign key in either the fact table or another dimension table.
For more information about cubes, see Defining Cubes.
You can also build virtual cubes, which combine elements of multiple, previously built cubes. When users browse the virtual cube, they see the combined elements together as if they were in a single cube. One of the main advantages of virtual cubes is that their definitions, but not their data, are stored. Thus, virtual cubes require extremely little storage space compared to ordinary cubes.