Extending Data Products

The data products that are provided by SAP are called SAP-managed data products. They are shipped for different levels of integration:

• Primary data products: Raw data (replicated from the original source, usually with a very basic level of refinement)
• Derived data products: Enriched data (with added semantics and additional fields)

When SAP ships an update, you will see in the SAP BDC Cockpit a clear indication that an update is available. However, you may still require further enrichment that go beyond the enhancements from SAP.

For example, you might want to include customer-specific fields. If your use case is not covered by an already available data product, you can first create your own data products, based on other existing objects:

• InfoObjects or InfoProviders from SAP BW, Private Cloud Edition, within SAP BDC.
• Entities, such as tables or views in the SAP Datasphere component of SAP BDC.
• Data sets from the open ecosystem of SAP BDC (for example, SAP Databricks, SAP Snowflake).

The new data products are called customer-managed data products. You can create customer-manged data products on different levels of integration, harmonization, and enrichment. For example, you can combine objects of SAP-managed data products with objects of customer-managed data products to combine data sets from your historic data warehouse, external database tables, ABAP CDS views, and machine learning algorithms. Then, you can enrich with calculations and semantics, and publish the derived data set as a new customer- manged data product.

Let's look at how such a scenario is implemented.

Individual data products can be directly consumed by applications and AI agents. But for analytical applications, individual data products usually do not provide sufficient semantics. Analytical applications require rich semantics that fully describe the data and how the data should be handled by the analytical tool. Although the data products do carry a basic level of meta data, much more analytic-related meta data is needed by analytical tools to enable business users to navigate their data across dimensions and hierarchies. Also, SAP data alone is sometimes not enough to build an analytical application. Analytical applications often require non-SAP data to be combined with SAP data to provide the big picture.

So how can you combine and enrich the data products, add additional meta data, and include third-party data so that you have a data set ready to be consumed by an analytical application? The answer is, by using the SAP Datasphere component of SAP BDC.

SAP Datasphere sits at the center of SAP BDC as a key component of a formation. Datasphere is the integration hub for the connectivity of SAP and non-SAP data sources and includes a powerful data modeling toolset.

Datasphere supports multiple data modeling approaches including relational and dimensional modelling. Datasphere includes tools to connect to any source system, data acquisition (either federated or replicated), data transformation, harmonization, and semantic enrichment.

To get started you first need to install one or more data products to a custom Datasphere space. The data products must already have been installed to the object store of the Foundation Services of SAP BDC before you can do this.

You then navigate to the catalog of the SAP Datasphere tenant of your SAP BDC formation where you can search for the required data products that you would like to install to your custom Datasphere space.

Once you have found the data products you would like to use, simply select the Install button.

During the installation you will be asked if you would like to replicate the data or choose a federation approach.

• Replication - Datasphere local tables are generated and also replication flows to load data into the local tables.
• Federated - A federation approach means that remote tables are created in Datasphere that remotely connect to the data in the data products in the object store, without a physical copy of the data being made in Datasphere. SQL on Files is used to query the data products directly which are stored in a data lake file format in the object store.

The decision between replicate or federate is usually based around performance considerations. You might choose the replication option if data volumes are high and/or latency is to be avoided. Or you might choose the federation option if data volumes are low and/or latency is not an issue.

During the installation of a data product to Datasphere you must provide the name of the target space where the Datasphere artifacts will be generated. Once installed, you will be able to list the generated artifacts by opening the Data Builder in SAP Datasphere.

Let's work through the steps to install a data product to a Datasphere space using a guided exercise.

Click to Install a Data Product to a Datasphere space

Once your data product is installed to a custom SAP Datasphere space, it can be consumed in custom analytic models.

The individual tables of a data product will appear in Datasphere and can be used as data sources to create relational and dimensional models. You can include custom code using SQL to generate the precise output you need.

You can create connections to non-SAP data sources using a wide variety of adapters and connectors and then combine the non-SAP data with the SAP-managed data products.

In your custom Datasphere space you also define Data Access Controls (DAC) to determine who should have access to the data.

Let's enrich our standard SAP-managed data product by combining it with a custom table that was generated in SAP Databricks using a guided exercise.

Click to Enrich a Data Product in a Datasphere space

As well as installing the data product from the Foundation Services to a custom Datasphere space, you can also share the data product with SAP Databricks, SAP Snowflake and other data platforms that use the same open standards used by SAP BDC. In this case, the delta share protocol is used to implement a zero-copy approach.

Sharing data products with SAP Databricks and other data platforms is covered in the follow-on course Introduction to the SAP Business Data Cloud Open Data Ecosystem.