Generally speaking, the data structures that make up SAP BW can get pretty complex. Data structures within the system are made up of a number of smaller, integrated parts. With each new element added or changed, the data structure would need to be updated to reflect that change. Underlying connection to objects lead to inherent complexity in modeling since each of these objects are parented to a lot of derived child objects that may or may not be affected by a single system change. In order to make modeling simpler and more efficient, the aim is to simplify these data structures as much as possible. In SAP, InfoObjects are our primary data structure and the connected structures may be InfoProviders, DataSources or Transformations leading to a massively complex web making up the basis of a single structure. Here is where we start to simplify.
Reducing the Complexity
in SAP BW on Hana, open ODS views alongside things like CompositeProviders allow us to visualize our modeling differently. This simplification carries over to maintaining the system and ensuring that is much easier to deal with. On the modeling front, we get to utilize some pretty efficient techniques such as satellites, transitive attributes and using joins with CompositeProviders. In combination these give us some very powerful tools which we can use to do modeling and prototyping at a much faster pace than traditional SAP allows.
Exploring the Idea of Satellites
Satellites mimic the idea of publicly accessible variables from other object-oriented programming languages. Basically, a satellite allows for different aspects of a single InfoObject to be represented differently in multiple views. This leads to a more efficient method of visualizing the functions of a particular InfoObject under different headings. In such a case, departments only need to access the aspect of the element they are working with, reducing data consumption yet not impacting the functionality of the department as a whole, since the relevant data is still used in processing. Satellites also mean that data security is preserved, even though the data objects can still be accessed by the departments that need them.
In the traditional SAP BW system, there exists the ability to explore navigational attributes (known as transitive attributes) after a successful ETL run. In the current SAP BW on Hana, attributes can be accessed directly. In this case, it means that modeling can follow a sort of snowflake schema, whereby we break down complex transitive attributes into simpler and simpler levels through the implementation of transformation tables. This has the added benefit of minimizing the amount of redundant data the system has stored. Transitive attributes are a bit like derived classes with an object-oriented programming syntax. While all the internal InfoProviders are accessible, they are also one big data structure as well.
Joins using CompositeProviders
Traditionally in SAP BW, it was possible to execute a union operator to join InfoProviders. The Join command on SAP BW on Hana gives added functionality through the use of CompositeProviders. This allows the joining of different InfoProviders which can then be used as the basis of a second join. By continuing to join data elements down the chain, the end result is a heavily nested data structure that incorporates the individual elements of each join with business plan writers. This is perfect for agile implementation and makes the system a lot cleaner since there are no redundant objects and all nested objects can be directly accounted for.
Changing how we Think about Modeling
Dynamic modeling is the aim, and with tools like transitive elements, joins to replace unions and satellites, the goal of dynamic modeling is already attainable. Dynamic modeling using these techniques makes for faster ETL runs, and more dynamic use of system resources. BW on Hana is more than just a faster database – it’s a faster, more efficient way of getting processing of large data modules done efficiently and effectively.