OLAP cube

An OLAP cube or data cube ( OLAP cube english or english data cube ), also called cube operator, is a common in the Data Warehouse theory concept to the logical representation of data. The data is arranged as elements of a multidimensional cube (English cube ). The dimensions of the cube describe the data and provide a simple way to access it. Data can be selected from one or more axes of the cube. The term OLAP (Online Analytical Processing ) is derived from the data analysis.

This type of representation for the analysis of data is advantageous since the data is accessed in the same manner on different aspects (dimensions). Hence the use in OLAP applications that analyze the data in a data warehouse or visually display.

Basic operations

  • Slicing: cutting out slices from the data cube
  • Dicing: Simultaneous slicing operations in different dimensions. Here, a smaller cube is created which contains a portion of the total cube.
  • Pivoting / rotation of rotating the data cube, so that at least one other dimension is visible
  • Drill- Down: Break down aggregations of an information object to detailed values; " Zooming in "
  • Drill-Up/Roll-Up: counter-operation to drill down; Compression to higher hierarchical level (eg, from the month of the year as a whole )
  • Drill- Across: dimension on the same hierarchical level; Consideration of the adjacent dimension members ( another region, another product, another month )
  • Drill-Through: refinement up to the highest level of detail (eg master data record, transaction receipt )
  • Split: The Split operator allows a value other dimensions further divide to reach further levels of detail (for example, sales of a store for a certain amount of products )
  • Merge / Drill-In: In contrast to split here, the granularity is reduced by the removal of extra dimensions again.

Example

OLAP cubes are often used in the analysis of corporate data is used, for example, sales, stocks and sales. Among the dimensions that may be important here include, for example, time, store, seller and product.

So the cube represents the data ( also called facts) sales, inventory, sales dar. depending on the dimensions of time, store, seller, customer and product

It can therefore be very easy to answer the following questions:

  • How much coffee was sold last week in the store Marburg?
  • How much coffee is there in the camp?
  • Which seller has sold the most coffee?
  • Which branch has last year made ​​the most revenue?

Technical implementation

The data is stored multidimensional (MOLAP ), relational ( ROLAP ) or in hybrid configuration ( HoLAP ). Some systems load the data in the initialization in main memory to allow fast access. In general, the cube is " sparse " (English sparse ), that is, the vast majority of possible intersections in the cube are not used with numbers. The handling of software with these parts of the cube contributes significantly to the memory requirement and the performance of these systems.

For relational systems use a star schema is typical. It is a separation into a fact table and several dimension tables clustered about made.

  • Business Intelligence
  • Data Warehousing
Data-Warehouse Data drilling
209328
de