Measures are the numeric data of primary interest to cube users. The measures you select depend on the types of information requested by users. Some common measures are sales, cost, expenditures, and production count.
Measures are aggregated by Microsoft® SQL Server™ OLAP Services, and the aggregations are stored for quick retrieval by users querying cubes. You select the measures for a cube when you build it with the Cube wizard or the Cube editor. You also select measures when you build a virtual cube with the Virtual Cube wizard.
Each measure is stored in a column in a fact table in the data warehouse. Because a cube can have only one fact table, all of a cube’s measures must be contained within it.
A measure can contain multiple columns combined in an expression. For example, the Profit measure is the difference of two numeric columns: Sales and Cost. For information about adding this type of measure to a cube, see Adding a Multiple-Column Measure to a Cube.
Fact table columns can be additive or nonadditive, and OLAP Services can use both types as measures. Additive columns can be summed. For example, a monetary column is additive. Additive columns are suitable as measures in a cube. Nonadditive columns cannot be summed meaningfully. For example, a numeric column containing an identifier such as Account Number is nonadditive. Nonadditive columns themselves are not suitable as measures in a cube, but they can be combined with some functions such as Count. The result can then be used as a measure. To apply a function to a nonadditive column so that the result can be used as a measure, create a calculated member.
Calculated members can be used as measures. Calculated member values are created from formulae when the cube is browsed, but the values are not stored. Thus, calculated members save storage. For more information, see Calculated Members.
Measures form the core of cube information presented to users. Presentation may be tabular or graphical depending on the tool with which users browse cubes, but the measures are the information users focus on.
In tabular presentations, measures are displayed in rows and columns. Whereas a cube’s dimensions determine the column and row headings, the measures are the data in the rows and columns. However, if you specify multiple measures in a cube, they too provide headings to separate the measures.
In graphical presentations, measures may be displayed in a variety of ways including lines, shapes, colors, shades, and shadows. Nevertheless, as in tabular presentations, the measures occupy the focal portion of the presentation while the dimensions provide peripheral labels.