Introducing Statistical Analysis


After completing this lesson, you will be able to:

  • Describe the new implementation of statistics

Introducing Statistical Analysis in SAP BW/4HANA

In the long history of SAP BW, the technical content (namespace 0TCT*) always played an important role. Useful statistics related to the core processes of the system, such as reporting runtimes, data provisioning processes, and aggregates processing were collected. The statistics were collected using DataSources that extracted data from the BW system itself (RSDDSTAT* tables) and loaded to InfoCubes. Some VirtualProviders (InfoCubes with DTP) were also included to collect real-time statistics. MultiProviders provided the interface for a huge number of delivered BEx queries.

Due to the type of objects used, this technical content is not compliant with SAP BW/4HANA. Instead of rebuilding the same application with DataStore Objects (advanced) and CompositeProviders, SAP BW/4HANA provides a pure virtual data model (VDM) based on SAP ABAP CDS views. The backend tables and their administration (transaction RSDDSTAT) are still the same, but the whole application architecture on top has changed to a pure virtual one. This solution is rolled out in SAP BW/4HANA support packages, rather than the Content Add-On.

The Query Monitor (Tr. RSRT) is a good starting point to explore these runtime statistics in SAP BW/4HANA.

Under the InfoArea SAP BW/4HANA Runtime Statistics [2O-BW4-DM_STAT], Transient Providers generated by the consumption CDS Views and related transient Queries on top of them are listed for evaluation of their content. This InfoArea is normally not visible in the BW Modeling Tools; however, it is still possible to define custom BW queries on these transient providers. When creating a new query, make sure you enhance the scope for Transient Providers so you are able to find them (technical names start with 2CRV*).

SAP BW/4HANA Statistics

The following SAP BW/4HANA statistics are currently available:

  • Query Runtime Statistics:

    • Query Runtime Statistics
    • Aggregation of Query Runtime Statistics
  • Process Chain Statistics:

    • Statistics for Process Chain Status
    • Statistics for Status and Runtime Information of Process Instances
  • Data Loading Statistics:

    • RSPM Request Statistics
    • RSPM DTP Load Statistics
  • Data Volume Statistics:

    • Statistics for the combined SAP HANA/cold store data volume
    • Statistics for SAP HANA online data volume
    • Statistics for cold Store data volume

For more details please refer to:

SAP BW/4HANA help: Statistical Analysis of the Data Warehouse Infrastructure:

Conversion Aspects

The biggest difference between the classic technical Content (0TCT*) and the new SAP BW/4HANA Statistics is that the new concept relies on a pure virtual data model and no data loading processes are required any more. In the past, the statistical data in scope was loaded regularly from the SAP backend tables into a fully BW-managed application. Afterward, the source tables were cleaned up in order to govern the data growth of them based on the RSDDSTAT_DATA_DELETE report (see SAP note 2470847).

As the new architecture does not manage any own data persistence, the cleanup of the source tables actually deletes the data which might be expected for reporting. For this reason, this housekeeping task should be re-evaluated and changed to much larger intervals: this means, while in the past the source tables kept only a history of two weeks for example, this retention period should be enhanced to a much larger time interval (for example, 24 months) to make sure you have all reporting-relevant data available.

Although being a virtual data model by origin, you still have the option to load the relevant data into an own customer data model, similar to the one of the past. The SAP BW/4HANA VDM provides some consumption views which are enabled extraction by default. These views support full extraction and can be used with ODP to model historical analysis by loading snapshots and subsequently creating time series. Based on an ODP-myself source system with context ODP_CDS, you can create DataSources and load the snapshots into your own DataStore Objects (advanced).

This approach might also be interesting for conversions, for instance where you convert the classic 0TCT-InfoCubes into DataStore Objects (advanced) and after conversion to SAP BW/4HANA, you use the new ODP_CDS DataSources to keep loading the statistical data in a similar way as you did it before without disrupting the analytical customer application on top of your existing data model.

For more details please refer to:

Log in to track your progress & complete quizzes