cancel
Showing results for 
Search instead for 
Did you mean: 
Read only

Datasphere CDS View live modelling

Tobias_Binkertm8
Explorer
0 Kudos
252

Within our SAP Datasphere we are interested in providing live insights and reducing data footprint (nothing unusual so far). However, therefore we need to import remote tables based on CDS Views.

Unfortunately, when trying to select typical CDS Views like I_JOURNALENTRYITEM which would be a perfect fit for live data based on remote tables these views are not displayed when chosing remote tables within our system as they miss required annotation.

 

For modelling purposes based on live views it would be easiest to have CDS Views which are based on the table only and are kind of a table wrapping CDS View for further modelling in Datasphere.

With the link below it is easy to identify the views that are based on an object (e.g. a table).

However, there is mostly no wrapping view that can be used in datasphere on the table directly.

Most likely it behaves the same for other customers as well as it should be a common use case that modelling based on live data is preferred over replication and modelling is easier when adressing a table or view that displays the tables cells.

Do you have any hints for me or is it really required that every customer creates his own CDS Views based on the table with the required annotations for live modelling?

Thank you very much in advance

View Entire Topic
michal_majer
Active Participant
0 Kudos

My suggestion would be to implement a two-tier approach: replicate standard CDS views via Replication Flows (supporting both initial load and delta synchronization) followed by creating semantic View wrappers on top of this imported raw data. This architecture balances near real-time data currency with optimized query performance while circumventing the annotation constraints of direct remote table consumption. The View wrapper layer also provides an abstraction that shields downstream models from potential structural changes in the source system.