Measures
In IBM® Cognos® Dynamic Cubes, you can define regular measures and calculated measures.
Regular measures are mapped directly to a database column of numerical data or defined by an expression. If defined by an expression, the expression is constructed from relational metadata and cannot include dimensional constructs and functions.
Calculated measures are computed in the context of a dynamic cube and computed in the dynamic query server. The expression is constructed from cube metadata and uses dimensional constructs and functions. Dimensional expressions are required when it is necessary to traverse hierarchical relationships or compute complex calculations which are difficult or impossible with relational expressions. With dimensional expressions, you have the ability to access parent/child relationships, to calculate parallel periods, to use set operations and to define an expressions which are evaluated based on its context within a query.
There are some behavior similarities between calculated measures and calculated members. For information about calculated members, see Calculated members.
In Cognos Dynamic Cubes, a measure dimension, containing a set of measures, is used in a dynamic cube as the center of a star schema. The physical grouping of measures into a single fact table implies that they share one area of interest. Each measure references the attributes that are used in measure-to-dimension joins. Each measure also references the attributes and joins that are used to map the additional measures across multiple database tables. The value of a measure is meaningful only within the context of the dimensions in a cube. For example, a revenue of 300 has no meaning on its own, but does have meaning in the context of dimensions, such as Region and Time. For example, the revenue for New York in January is 300. Common examples of measures are Revenue, Cost, and Profit.
Simple arithmetic expressions can often be evaluated either by the relational database or in the context of the cube. If a measure expression can be evaluated in either context, it may be preferable to choose a relational expression. Relational databases usually have access to wider range of functions and may be more efficient. If a database is constrained in terms of resources, an alternative is to use calculated measures.