SAP BW data integration

SAP BW data integration

Why is data analyzed outside SAP?

PowerBI/Tableau is very limited when connecting to SAP BW data

Databricks solution: PowerBI/Tableau provides native and performant connectors to Databricks

non-SAP data are not available in SAP BW and cannot be easily added there

Databricks solution: any data can be loaded into Databricks easily and combined with SAP data

SAP BW development is expensive and time-consuming (strict rules applied)

Databricks solution: finish your analyses faster due to less strict rules applied

skilled SAP consultants are a scarce resource

Databricks solution: simple SQL can be used to analyze the data

SAP BW can be slow when too many users analyze data at the same time

Databricks solution: Databricks can scale up and down on-request

SAP BW does not provide tools for predictive analytics/machine learning

Databricks solution: various tools and libraries to build predictive algorithms and ML solutions

Our Accelerators


Authorization integration

Sync your SAP BW and Databricks data access permissions.

Incremental data load support

For BW InfoCubes or Composite Providers.

BW query conversion

Move your BW query logic to Databricks automatically to prevent logic duplication.

Metadata-driven framework

For accelerated and governed reporting layer development

Cut SAP BW costs by using Databricks

By copying your data into Databricks, you may:

  • avoid buying more RAM or servers in general by limiting the amount of data stored in SAP HANA
  • lower your SAP maintenance and support bills

SAP BW Data Ingestion Architecture

Ingest SAP objects from classic SAP BW or SAP BW/4HANA into Databricks to build a data model by combining SAP data with some non-SAP data.


Data extraction best-practice (ELT > ETL)

  • extract all columns to let the data analytics/scientists decide which data they need
  • we do not recommend using S/4HANA CDSViews as it creates a dependency on the SAP S/4HANA team
    • developing CDSViews using the standard S/4HANA release process could be very time-consuming
    • SAP-built CDSViews change quite often, we cannot control that (physical tables do not change that often)

Expected effort/budget needed:

  • ~100 - 200 MDs for the initial setup (Databricks platform setup, BW data ingestion, data model in Databricks, authorization integration):
  • 1 - 5 MDs per each SAP object (Multi Provider, InfoCube, CompositeProvider, aDSO)
  • €80k - €160k per year for data ingestion tool fees, Databricks compute, cloud storage and I/Os