extract logic/tables/sequence?

[Former Member]

Hello,

I'm curious about the sequence in which different kinds of extraction take place leading to data loading into the BW system (either master- or trans. data).

Is this a correct assumtion:

1. Masterdata: during execution of the info package, the function module (alternatively, a view is used as source and no function module is needed) of the called data source goes into the source tables (ex. MARA), returns the data via the rules of the extract structure (incl. appends/user exits).

2. Trans. data init:during execution of the info package, the function module (alternatively, a view is used as source and no function module is needed) of the called data source goes into the source tables (GLPCA), returns the data via the rules of the extract structure (incl. appends/user exits). After the load, a delta table is generated. Within f.e. SD, you have to generate and fill a setup table in order to be able to load th init.

3. Trans data delta: new postings are added to the delta table (in PCA this happens automatically while in f.e. SD, you have to schedule this procedure) from which the new postings are fetched (via function module or view) on execution of the info package in BW.

Is this the way it works in general? Are there any good documents on this topic? Please mail to fc.sapbi@gmail.com

Regards,

F C

Message was edited by:

F C

[Former Member · Accepted solution]

1. For master data like material, the exractors behave all most the same way.

2. For transaction data, it is diferent for LO comapred to non-LO extractors. for LO transaction data, there are some extra steps you need to do like fillng the set up tables, choosing a delta method, V1,V2, V3.

3. For delta, like I said in 2 above, it is slighly different for LO extractors.

There is lots of info on this subject in SDN. Search with key word LO or cockpit will be very useful.

Ravi Thothadri

[Former Member · Accepted solution]

Hello,

What I'd like to understand more about is the exact logic of the extraction process, not the actual steps for the end-user.. E.g. which tables are used during the process. When does the function module come into the picture. Which is the relationship between the FM and the extract structure. Where is the extracted data stores etc. etc.

Is there any good documentaion on the subject?

Regards,

F C