Hypercube OLAP

A hypercube OLAP (or cubic OLAP ) is an abstract representation of multidimensional information exclusively numerical used by the approach OLAP (acronym of On-line Analytical Processing ). This structure is provided at ends of analyzes interactive S by one or more people (often neither data processing specialists nor statisticians) of the trade that these Donnée S is supposed to represent.

Cubes OLAP have the following characteristics:

  • to obtain already aggregate information according to the needs for the utilisateur.
  • simplicity and speed of accès
  • capacity to handle the Donnée S incorporated according to various dimensions
  • a cube uses the traditional functions of aggregation: min, max, count, sum, avg, but can use functions of aggregations spécifiques

Primitive

Hypercube OLAP gives access to functions of extraction of information (for visualization, analyzes or treatment), and to functions of request in language MDX (comparable with SQL for a Database relational).

  • Rotate : selection of the couple of size to be targeted,

  • Slicing : extraction of a section of information,
  • Scoping : extraction of a block of Given S (operation more general than the slicing ),
  • Seed-planting drill-up : synthesis of information according to a dimension (example of seed-planting drill-up on the axis time: to pass from the presentation of information day per day over one year, with a synthetic value for the year),
  • Seed-planting drill-down : it is the equivalent of a “zoom”, opposite operation of the seed-planting drill-up ,
  • Seed-planting drill-through : when one has only Donnée S aggregate (added up indicators), the seed-planting drill through gives access the elementary detail information (see in particular the tools H-OLAP).

Implementation

The concept of hypercube OLAP can be used by the means of various interfaces according to the needs and the capacities. There are currently four.

  • M-OLAP : The most traditional form because fastest. It uses multidimensional tables to save information and to carry out the operations.
  • R-OLAP : That which requires less investment. She works on relational tables. A new table is created to contain each aggregate.
  • H-OLAP ( Hybrid OLAP ): She uses at the same time the relational tables to store rough information, and of the multidimensional tables for the aggregates of predictive information.
  • D-OLAP ( Dynamic or Desktop OLAP)

See also the list of decision-making tools.

Use

In the beginning used for an instinctive analysis of information, hypercubes OLAP can be coupled with systems dated mining and thus to analyze, predict and simulate more “strictly” information.

The companies generate this type of structure by massive synchronizations of information (LTE) starting from basic management systems of data relational (DBMS R) of type Datamart , Datawarehouse or sometimes even compromise according to the architecture which they chose for their decisional Information system.

The restitution of information is done using simple requêteurs using the language MDX ( Multidimensional Expressions ), of Executive Information Systems (EIS), of specialized applications (software of company), or in a Tableur (provided with a Plugin of navigation).

Example

As illustration, it is possible to analyze the sales turnover of a company in 4 following dimensions:

  • Geography: continent > country > area > department > city
  • Time: year > quarter > month > week > day
  • Product range: range > standard > family > reference
  • Organization: department > sector > responsible > salesman

It is possible to find furnished documents on language MDX on certain sites presenting of documentation on line. Site MSDN France gives good bases of information relating to this language: .

See too

Random links:Gift Simpson | Amédée Pierre | Feature (linguistic) | Camille Bloch (historian) | Michigan City (Indiana)