Question:
Hello All
I have just started to work on implementing some delta extractors
can any of you please clarify what is all about delta queues,etc.
I have done lots of generic extractors all full extractors.
How does deltas in SAP BW work I have seen from RSA7 there is delta
queue not sure how these delta extractors work and where the delta
queue plays a role can any of you kindly explain this to me. Also, what role does ODS play in the deltas I see some of the AR/AP scenarious have a ODS for the deltas how does the ODS and deltas interact. Is the ODS used to extract deltas please explain
Thanks
Tracey
Answer:
Simply..
Deltas give you just the changed/new records since your last load. Once you do a initialization run from BW for a delta enabled datasource an entry appears in RSA7 for that datasource. .
ODS can be used to transfer deltas to a cube when the datasource does not provide deltas (like the ones you wrote)
Answer:
Thanks for your inputs.
Does it mean that wherever a ODS is used a full extract is done from R/3 and BW compares using the ODS the current request and the previous request what as changed and loads only those records that are changed into the cube. So that means with a ODS still all data is extracted from R/3 and the cubes get only the incremental delta.
I am still not clear what the delta queue does and its purpose. Does the delta queue maintain only the dates on which the last extract was run and does BW use this to get the cutoff dates to extract the dates to extract from R/3.
Does it not mean that delta queues can work only when the tables providing data have a last changed date in them otherwise this might not be useful am I correct.
Can you please kindly elaborate I am new to delta extracts it is so difficult understanding how they work.
Thanks
Tracey
Answer:
Delta processing is described in some detail in the SAP documentation for ODS (V2.0). There is a White Paper on the ODS, and a longer document in the How To.. Guides on sapnet (//service.sap.com/bw).
Since an InfoCube contains aggregated data the challenge of delta processing is to extract all new transactions and changes without duplication or gaps.
The ODS has a flat table structure with a transaction key that can be used to compare a new set of transactions with the existing ones. Through this comparison the ODS delta logic is able to generate the InfoCube update that reverses old values and adds new ones, and ignores duplicates.
For efficiency when using this method send as few records as possible (NOT a full extract) to the ODS to be sure of including all new and changed transactions since the last data upload.
Example:
The active data of the ODS includes a transaction that has been loaded into an InfoCube:
Key=X, Amount=$100, Items=1
The new/modified data includes a revision for this transaction:
Key=X, Amount=$80, Items=1
As a result of this change the following transactions will be created to be loaded into the cube:
Key=X, Amount=(-$100), Items=(-1)
Key=X, Amount=$80, Items=1
.. the first reverses the previously loaded transaction, including the transaction count (items); the second adds the new values.