cancel
Showing results for 
Search instead for 
Did you mean: 

Modeling performance...

Former Member
0 Kudos

I'm building a model, and would like to build it in smaller blocks.  That is, I'd like to build an AT view that represents Billing (VRBK, VRBP), and another that represents shipping/deliveries (LIPS/LIKP/VEPO/VEKP).  When I run a native SQL query screening on Invoice Number (VBRK.VBELN), the query works perfectly.  However, if I join the two AT's in the same manner in which the query is written, performance is crazy, as the execution plan indicates that the shipping AT is bringing back 31M records.  How can I build this such that I can continue to build modularly, while getting the performance I need?

Accepted Solutions (0)

Answers (2)

Answers (2)

KMS71
Participant
0 Kudos

Hi,

As Sergio mentioned, there could be other ways that could provide a near standard solution Have a look at HANA Live and the delivered queries...which you could use as templates, or even extend the delivered views if appropriate. This would be much easier than starting from scratch...

Thanks

Kris

SergioG_TX
Active Contributor
0 Kudos

Peter,

if you are just starting with your modeling.. i suggest you use calculation views as per SAP direction.

as far as how to improve performance... there may be general db techniques such as filtering low, aggregating on top, etc.. add filters to minimize data returned in the modeling side.. further, you will also benefit by parallel execution depending on how you are modeling... avoid branches of code, etc.

there are many ways to help improve performance, however, it will depend on your business scenario as there isnt a one fixes all approach. I hope others may share additional details .