SQL consumption of BW on HANA data models
This document will focus on the HANA model generation based on BW data models. This is a feature that was already introduced with BW 7.3 on HANA especially for SAP BOBJ Explorer. More information you can find in that specific document. https://scn.sap.com/docs/DOC-59791
This page here will focus on the features of that integration architecture independent of BI clients.
What is the purpose and positioning of the generation of HANA models based on BW models as it allows native HANA reporting on BW data?
See the following link with a very good explanation of the intention of this feature and what it shouldn't be used for.
In BW 7.4 a new HANA model generation process was introduced, also called 'Push' approach. What is the difference to the previous 'Pull' approach?
Both approached are described in detail in a First Guidance paper - https://scn.sap.com/docs/DOC-52790
In the SAP Help Portal you can also find more information - http://help.sap.com/saphelp_nw74/helpdata/en/66/33d851345c4770bd4e523701b9f5b0/content.htm
The main difference is that the previous Pull-approach as mainly described on the BOBJ Explorer FAQ page (https://scn.sap.com/docs/DOC-59791) requires a HANA modeler to go into HANA Studio and manually import a BW model. If changes are made on the BW side to the corresponding model (e.g. InfoProvider model change) this generated HANA view would most likely not work anymore and needed to be re-generated. This is a manual process as well.
The ne Push-approach manages all those steps on the BW side. How this works is that the corresponding BW model (e.g. an InfoProvider) is flagged to automatically generate a HANA model during creation. That way the entire HANA model generation process becomes part of the BW lifecycle management including transport management, authorization integration etc. It will also check if a change to the BW model would require an automatic re-generation of that HANA view which might have an impact on the additional HANA modeling that has been put on top of that view via HANA Studio.
Click on the following link to see a blog about the new Push-approach and more details.
Will the new push-approach replace the previous pull-approach?
No. There is a very valid use case for both approaches.
What is SAP's recommendation about when to use which approach?
If the intention of the HANA model generation is to enable SQL consumption of BW data and reporting via discovery and exploration BI clients (e.g. SAP BOBJ Explorer and SAP Lumira) the new push-approach is probably the best option as it comes with the highest integration to the BW management and governance layer, e.g. authorizations, transport management etc.
However, if the purpose of the HANA model generation is to integrate BW data with other data in HANA that is managed outside BW, e.g. to build a data mart directly with HANA Studio, where the BW data is just one part of the entire data model, then it's probably better to use the pull-approach, since authorizations and the entire management for the entire data mart data model is done via HANA Studio.
Is it also possible to generate HANA Views based on BW BEx queries?
Since BW 7.4 SP8 it is possible to consume a BEx query via an automatically generated HANA View. See the following link for more details and the currently supported BEx features. Important to mention here is that those generated HANA Views based on a BEx query are also able to read SAP IQ NLS data!