Dimension Hierarchies

Analysis Services

Analysis Services

Dimension Hierarchies

A hierarchy is the set of members in a dimension and their positions relative to one another. For a dimension created from a data mining model, the hierarchy represents the node structure of the mining model.

Hierarchies are sometimes represented as pyramidal structures. The only exceptions are hierarchies in which all members are at the same level. An example is the hierarchy of the Measures dimension.

From the top of a pyramidal hierarchy to the bottom, the members are progressively more detailed. For example, in a Geography dimension defined with the levels Continent, Country, and City, in that order, the member Europe appears in the top level of the hierarchy, the member France appears in the middle level, and the member Paris appears in the bottom level. France is more specific than Europe, and Paris is more specific than France.

The lower the level of a pyramidal hierarchy, the more members it usually contains. In the preceding example, there are more countries than continents and more cities than countries.

Microsoft® SQL Server™ 2000 Analysis Services supports several types of hierarchies: balanced, unbalanced, and ragged.

Balanced and Unbalanced Hierarchies

In a balanced hierarchy, all branches of the hierarchy descend to the same level, and each member's logical parent is the level immediately above the member. In an unbalanced hierarchy, branches of the hierarchy descend to different levels.

For more information about balanced and unbalanced hierarchies, see Balanced and Unbalanced Hierarchies.

Ragged Hierarchies

In a ragged hierarchy, at least one member's logical parent is not in the level immediately above the member. This can cause branches of the hierarchy to descend to different levels.

For more information about ragged hierarchies, see Ragged Hierarchies.

Dimensions with Multiple Hierarchies

Analysis Services supports dimensions with multiple hierarchies. These dimensions provide similar yet alternate views of cube data. For example, a Time dimension with two hierarchies  can have a regular calendar hierarchy and a fiscal calendar hierarchy. In Analysis Services, a dimension with multiple hierarchies is defined as two or more dimensions with names that share the same prefix followed by a period but have different suffixes (for example, dimensions with names of Time.Calendar and Time.Fiscal). The suffix should not equal any current or future level name or member name in the dimension because queries using the dimension may be ambiguous.

See Also

Creating Dimensions with Multiple Hierarchies

Data Mining Models

Dependent Dimensions

Parent-Child Dimensions

Ragged Hierarchies

Ragged Dimension Support

Regular Dimensions

Virtual Dimensions