Question:
Dear All,
We would like to go for FICO data extraction from SAP r/3 to BIW. Our data size is very very high (Approximate `~300,00000 Records/ODS).
I would like to know, What strategy we should adopt for FICO Extraction. Should we lock all the production FICO user at the time of extraction or any other alternative available.
This is very very urgent. Please guide us for the same.
Thanks & Regrads,
PBA.
Answer:
We would like to go for FICO data extraction from SAP r/3 to BIW. Our data size is very very high (Approximate `~300,00000 Records/ODS).
Do not worry I managed to load similar amount of data before for CO Line Items. I created the DataSource and extractor that reads from view COVP in R/3 and extract CO line items into BW. Beware there is no SAP delivered DataSource for CO Line Items (my knowledge is up to BW2.0b).
I would like to know, What strategy we should adopt for FICO Extraction. Should we lock all the production FICO user at the time of extraction or any other alternative available.
You will need to segregate your data load into multiple extracts. You need to plan it very well before you actually start loading. You will not need to lock your CO users at all.
In order to help you more, provide more information as follows:
1. What is the DataSource you are going to use?
2. If you are going to build this DataSource, which tables you are going to read from in R/3? Is it COVP view in R/3?
3. Are you requiring only one controlling currency for your reporting? Do you need to report on Controlling Area Currency, Company Code Currency, and Document Currency as well? Indicate what currency you would like to report on.
Regards,
_________________
Ali H. Alwan
Atos Origin Middle East
Saudi Arabia
Answer:
Thanks ali,
Actually we are using standard extractors of PCA. They are:
0EC_PCA_1
0EC_PCA_3
Ya, Please explain me the detail procedure for deviding load into multiple extract. Specially with regard to non user locking.
We are actully planning 75,0000 Records per Load.
Also tell me for every upload we have to take data with initialize delta.
Regards,
PBA
Answer:
Ya, Please explain me the detail procedure for deviding load into multiple extract. Specially with regard to non user locking.
We are actully planning 75,0000 Records per Load.
You simply need to restrict your data load in InfoPackage selection tab. You need to do different data selection for every data load. By this you divide the data load into multiple extracts.
Choose carefully the fields you are going to restrict the data load based on. Do some analysis of these fields in R/3 to determine how many records expected for each data restriction. If you know the tables that these two DataSources are reading from then you can try different values of your selection fields and see how many records hit they drive.
You will not need to lock your users at all by making sure that each data extracts are relatively small for the system to handle smoothly. For 30 Million records you mentioned you can try loading roughly 1.5 Million for each extract (try to hit for this number). Therefore you are going to do 20 data loads (delta-initializations)
Also tell me for every upload we have to take data with initialize delta.
Yes you have to do delta-initialization for each data load. Until you complete loading all existing data in R/3. Then you carry on with subsequent deltas as usual.
Regards,
_________________
Ali H. Alwan
Atos Origin Middle East
Saudi Arabia
Answer:
Hi ali,
I have scheduled the job and took all data into the psa.
I am taking data into the ods. Can I schedule the job parallel processing or i have to go for serially??
Parallel process can affect my data quality??
But I am maintaining seriallization in Job activation.
this is very urgent. What is your opinion.
PBA