Dimensions are categories used to summarize the data of interest. Dimensions created by a wizard generally prove to be good first drafts but need refinement.
Chapter 51 presents background on Business Intelligence and data warehousing concepts.
Dimensions contain attributes, which are the way data in the cube will be sliced. For instance, a cube may contain a measure group of sales that stores the product sold, the sale amount, and the date of the sale. A user may want to see sales totals by year and would slice the data by calendar year, an attribute or characteristic of the date dimension.
In addition to storing attributes, a dimension must have exactly one key column. The key column denotes how to uniquely identify each record. A date dimension may have quarters, months, and years, but the data key uniquely identifies each record. In this case the date key could be the concatenation of month, day, and year.
If you do not create dimensions during the Cube Wizard process, you can manually create them independent of the cube. Before creating additional dimensions, be sure to add the tables required to build the dimension to the data source view. When all the required tables are in the DSV, right-click the Dimensions folder in the Solution Explorer, and select New Dimension to launch the Dimension Wizard.