07-24-2016 4:50 PM
Hello,
Is there any limitation regarding number of joins in an AMDP? For my requirement, i must join 32 tables. Is it a performance issue to join so many tables? I can create two or three AMDP's as well but i read in one of the post that the transfer of data must be minimal or else it could affect the performance. Can you suggest me an approach?
Thanks,
Prem
07-25-2016 11:17 AM
Generally speaking every additional join will incur additional processing, which means: this likely will take more time.
The sheer number of tables here is not as relevant as the amount of data processed with every join.
SAP HANA Live virtual data models, as an example, commonly have tens of stacked views and joins.
As the SAP HANA query optimiser tries to find the best possible way to execute such joins, trying to implement them manually very likely won't be a good idea.
So rather create views to logically put joins together that belong together, but don't artificially build multiple ADMPs.
07-25-2016 7:09 AM
hi,
Can't you create CDS for the same and use the result ?
Regards,
Naveen Kumar.
07-25-2016 11:17 AM
Generally speaking every additional join will incur additional processing, which means: this likely will take more time.
The sheer number of tables here is not as relevant as the amount of data processed with every join.
SAP HANA Live virtual data models, as an example, commonly have tens of stacked views and joins.
As the SAP HANA query optimiser tries to find the best possible way to execute such joins, trying to implement them manually very likely won't be a good idea.
So rather create views to logically put joins together that belong together, but don't artificially build multiple ADMPs.
07-25-2016 2:36 PM
Hi Prem,
Create CDS View either create 2 or 3 graphical views and use it in AMDP.
If you really worried about performance, persist it in table with delta update later you can use it.
Have you heard about AMDP inheritance either calling AMDP inside AMDP?
-Amol S
07-25-2016 4:29 PM