on 07-20-2005 8:33 PM
Hai all,
We have a cube in FI. The cube is being loaded with a full load from an ODS. The option of "delete overlapping requests" hasn't been selected in the data target tab of info package. My question is aren't the records getting duplicated every time a full load is done? When I look at the contents of the cube, there are too many records for the one record. I understand that a cube donot addup the records if they are from different packets.but even then having those many records is wrong? Also we are reporting on a particular time period and we are getting the results ok.
My question is do I need to select the option " delete overlapping requests in info cube" or not?
Thank you.
Hi Visu,
For a full load the logic is simple -
If the same combination of master data(characteristic) exists then the key figures would be added else not.
For eg:
Emp-Name Dept Calmonth Salary
XXX 1 April 10000
XXX 1 April 10000
Would display in the report as
XXX 1 April 20000
whereas if the data is like this
Emp-Name Dept Calmonth Salary
XXX 1 April 10000
XXX 1 May 10000
Then it would display as 2 records
XXX 1 April 10000
XXX 1 May 10000
Once you get this logic clear then you can get things sorted out.
Hope it helps.
Bye
Dinesh.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
It depends on your dataflow, and its not mandatory that you have to delete the requests from the Infocube.
There could be some logic in the infopackage (Selection options) which only selectively updates the data in the infocube. OR, ODS could have been maintained such a way that it only contains the latest data (previous data could be deleted). These might be taking care of NOT updating duplicate data.
So, unless we know the complete data flow logic, there is no standard answer.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Visu,
If it is a FULL LOAD, definitely there would be duplication of data from ODS. I think you should consider dropping of requests in a FULL LOAD scenario or implementing DELTA.
In you update rules of the cube, do you have the option of Addition or No Update checked.
if it is No Update, the report might be having a parameter of considering only the most current request. This is possible by using variable of REQUEST ID like "Most Current Data" which is an SAP Exit. In this case, if the cube is compressed the entire data would report as duplicate in your case.
It is always desirable that the cube contains non duplicate data for efficient processing.
Regards
GPK
Thanks a lot Gova Mabbu.I cannot explain you the whole logic. But I will try my level best to get you an idea. This is how the data flow is:
From To Update Frequency
R/3 to PSA Daily
PSA to ZODS1 Delta update Daily
ZODS1 to ZODS2 Delta update EVERY 15 DAYS
ZODS2 to ZMCUBE Full update Same day after above delta
There is a logic that flags and deletes some unwanted records but I don't think it has anything to do with data loads. <b>Is there anyway to check if I have duplicate requests in an info cube?</b>I saw a record in PSA and when I opened the infoCube contents, the same record has about 28 other records with same values. Does that mean they are duplicated?
Thanks.
Since you are loading the data from ODS to InfoCube, your ODS data should match with the INfoCube totals, then you can be sure that there is no duplicate data coming into InfoCube from ODS... You can do this to check the data integrity. don't compare with PSA. compare ODS and Cube.
Also check if there is any logic in the start routines or in the InfoPackage selection conditions that takes care of not updaing duplicate data.
User | Count |
---|---|
84 | |
23 | |
11 | |
9 | |
8 | |
5 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.