With SAP HANA, there has been a paradigm shift in the way applications are developed at SAP. Core data services (CDS) simplify and harmonize the way to define, extend, and consume data models. ABAP CDS allows the definition of semantically rich data models providing all metadata that are necessary to build analytical or transactional apps or even remote APIs.
Modeling with ABAP CDS can result in SQL statements with a complexity that far exceeds that of the classic views of the ABAP Dictionary or of Open SQL statements in previous ABAP programs. Database accesses can become even more complex at runtime by tables joined to realize authority checks or when associations are followed.
The increased complexity makes it more difficult to predict SQL performance. If not implemented carefully, the response times of these datacentric applications can be very sensitve to the processed data volume and the selection criteria applied.
The session S4H300 – Boost Performance for CDS Views on SAP HANA at TechEd 2017 was a first heads up on what should be respected when using and implementing ABAP CDS views. Also for customers and SAP partners it is recommended to classify CDS views according to service qualities and access them appropriately. CDS view complexity should be kept to a minimum and calculated fields in ON-clauses of joins or key definitions should be avoided. Realistic tests with production-like data is a must-have.