Skip Headers
Previous
Previous
 
Next
Next

About Pivot Cubes

A pivot cube is the same as any other cube except that it has a measure dimension. A measure dimension has the names of the measures of one cube as its values. A measure dimension enables you to reference all the measures of a cube as a single entity, which simplifies the process of generating calculated measures. For example, if you have a cube with 100 measures, you can create a particular calculated measure for all of them in one procedure instead of 100 procedures.

To support a pivot cube, you first create a measure view from an existing fact table. The measure view has the same dimension key columns as the fact table, but two new columns, named MEASURE_DIM and MEASURE_VALUE, replace the fact columns. You create a separate measure dimension table for these columns, which provide the measures as foreign keys.

Table: Measure Dimension Table Description describes a measure dimension table.

Measure Dimension Table Description

Column Name Data Type Sample Values

MEASURE_ID

VARCHAR2(60 BYTE)

UNITS, SALES, COST

MEASURE_DESC

VARCHAR2(120 BYTE)

Units, Sales, Cost


You map the measure dimension to a measure dimension table, and the pivot cube to the measure view. When you create calculated measures, you can specify the measure dimension (represented in the cube as one measure) instead of multiple, individual measures.