Wednesday, 11 December 2013

Infocube

An InfoCube describes (from an analysis point of view) a self-contained dataset, for example, for a business-orientated area. You analyze this dataset in a BEx query.
InfoCubes are the central multidimensional data model in BI. Reports and analyses are based on InfoCubes. An InfoCube describes a self-enclosed data set encompassing one or more related business processes. A reporting user can define or execute queries against an InfoCube.
The following InfoCube types exist in BI:
  • InfoCubes
  • VirtualProviders 
Only InfoCubes physically contain data in the database. By doing so, they are also data targets, as data can be loaded into them. In contrast, VirtualProviders only represent logical views of a dataset. There is no difference between these InfoCube types as far as the reporting user is concerned. Queries can be defined based on all InfoCube types. InfoCubes are thus InfoProviders, as BI objects are called InfoProviders when queries can be defined/executed based on them in enterprise reporting.

There are two subtypes of InfoCubes: Standard, and Real-Time. Although both have an extended star schema design, Real-Time InfoCubes (previously called Transactional InfoCubes) are optimized for direct update, and do not need to use the ETL process. Real-Time InfoCubes are almost exclusively used in the BI Integrated Planning tool set. All BI InfoCubes consists of a quantity of relational tables arranged together in a star schema.

Note: A Real-Time InfoCube is a special InfoCube, specially developed for BI Integrated Planning or the older planning tool, BW /SEM (Strategic Enterprise Management) Business Planning and Simulation. Real-Time cubes were previously called Transactional InfoCubes, as the system accesses data in such an InfoCube transactionally. In other words, data is written to the InfoCube (possibly from more than one user at the same time) and instantaneously read again when required. Standard InfoCubes are not suitable here. Use Standard InfoCubes for pure read access

  • Fact table
A InfoCube consists of precisely one fact table* in which key figure values are stored. A fact table can contains a maximum of 233 key figures.
*Functional perspective
  • Dimension table
A InfoCube usually has a minimum of four dimension tables and a maximum of 16. Of these, 13 of the 16 are customer-created and three are the SAP-supplied dimensions:
  • Units dimension table
  • Data Package dimension table
  • Time dimension table
  • Customer dimensions contain SIDs linked to a maximum of 248 characteristics InfoObjects.
Hint: If you have a InfoCube with 13 DIMs x 248characteristics/DIM, you either have a weird business process or bad advice!
  • Data Package and Time dimension tables are always present in a InfoCube.
  • The Units dimension table only exists if at least one key figure is of type amount quantity.. In this case, a fixed/variable unit/currency needs to be entered with the key figure.
  • Dimension tables do not contain the characteristics/characteristic values, but the corresponding SID keys/values.
 
BI Extended Star Schema: Cost Center Transactions:


In extented star schema:
we have a intermediate table between master data table and dimension table..which is known as surrogate id table.

* We can have 248 x 13 dimensions... thus overcoming the limitation of dimensions in star schema...
* Master data is placed outside the cube.. so it can be reused...
* We don't have any alpha numeric characters inside the cube.. which will improve the performance of the cube.

About Star Schema:

  In cube we have fact table and dimension table...

In star schema,
                      master table will be dimension table
                      transaction table will be fact table.
 * fact table can hold only 16 characteristics.. out of which 3 is reserved for datapackage, time and unit..

    In short, we can have only 13 dimensions possible in star schema..

 *  As master data is used inside the cube .. no reuse of master data..

 *  Material is stored as "M01" customer as " C01"...

     so fact table must hold alpha numeric values.... which will degrade the performance of cube...


No comments:

Post a Comment