Using OBA for 2 sources (Siebel and EBS)

Hi,
We have OBIA for EBS - Supply chain implemented (with some custom) and it's already on Production. Now we are implementing OBIA for Siebel.
So, for the first time, we are planning run a FULL extract from Siebel. Then we noticed that both system uses some commom table (like W_PRODUCT_D). Unfortunally, we have on EBS all price list and products. On Siebel we have same information, replicated.
If we run a FULL extract, we'll lose the data from EBS products, I guess. But if we run a Incremental extract, I think we will have duplicate products on W_PRODUCT_D. Is it correct?
How is the best way to work it?
Any help will be great!
Regards,

Hi Ohara,
Regarding the RPD:
I am not completely sure if you need to change the RPD..
If you can guarantee that users will always report on a combination of dimensions with facts, there is no need to filter on datasource_num. This is because all facts loaded from Siebel will be related to dimension values loaded from Siebel and all facts loaded from eBS will be related to dimension values loaded from eBS. At least, this is how it is supposed to work, please allocate enough time to test this really good.
However, 'just to be sure' you can create business model filters in the security groups in such a way that all 'Siebel-users' will only see the Siebel data and vice versa.
Regarding the DAC:
I think the most ideal way would be to combine all the to-be-loaded Subject Areas in one Execution Plan. This can be done if extraction from both eBS and Siebel can be performed in the same timeslot. This way, the staging tables will be populated from 1 or 2 sources, but the SILOS mappings will only have to be run once.
Regarding the most interesting stuff (scenario 1 ;) ):
You will require an unique identifier of a Product/other that is present in both source systems to be able to match them in the BAW. Please do not resort to literal comparissons on names etc. Once you have this identifier, you can modify the LKP operators in the fact mapplets and/or reusable LKP operators to use this identifier to retrieve the wid from the dimension, instead of the integration_id and datasource_num.
Good luck!
Regards,
Marco
Edited by: m.siliakus on Feb 22, 2011 8:58 AM

Similar Messages

Maybe you are looking for