Showing results for 
Search instead for 
Did you mean: 

Errors in reports

Former Member
0 Kudos


I am working with FI-AR data thats coming from two different R/3 systems(say I and II). The data is extracted from datsources from R/3(each R/3 has different datasource) and finally consolidated in a ODS, a special program is run to check some conditions and delete unwanted records(as per the criterion) and then loaded into a cube.Once the data is in ODS, from then everything is same for both the data till the end(like the objects, change routines). My problem is when I am running reports, the results with data from the system I is perfect but there are errors in results from system II. Can anybody suggest me what the probable reasons are and how to solve this problem?

I am thinking of the following:

1.Extractor or datasource is not picking up properly for II?

2. Delta update is not catching up?

3. Extractors in R/3 needs to be checked for consisteny?

4. Initial data loads problem??

I appreciate any kind of help.

Accepted Solutions (0)

Answers (2)

Answers (2)

Active Contributor
0 Kudos


what is the error , I assume that:

1. You have no errors in data loading

2. The result is not what you expect .

What could be wrong is :

1. CHeck your update rules for overwrite / Addition , in most cases that might be the culprit since there is not too much clarity in how records are stored in R/3 - as to whether they are overwritten or stored in a different record when changed.

2. Check the logic of your transformation with an end to end use case , that would solve all your problems.

I would suggest you put effort into an end to end use case and try to validate it at every step.


Former Member
0 Kudos


We too had a similar kind of problem.

What we did is got the erroneous records from the report and with a top down approach traced those records from cube then ODS and so on..

Our error lied in the ODS due to some inconsistencies in data.

You might try doing the same..

Hope this helps..


Former Member
0 Kudos

Hai Ashish,

Thanks a lot for the reply. But the data we are reporting is not direct. I mean we have some pointers and a lot of structures involved.So we can't actually go down directly as there are too many combinations. Also the no. of records are more than 100. Do we have any other possiblity?

Thanks again.

Former Member
0 Kudos

Sombody please help!!!!!!!!

This issue is very important for me. ANy help is highly appreciated.

Former Member
0 Kudos


Its very difficult to say anything without knowing the data model..

I would suggest ..if your data is comming from multiple sources...get one example of erroneous record..say a particular doc and trace it from the extractor to the cube..

You could also do a test load for a very small dataset..

Good Luck..


Former Member
0 Kudos

Thanks Ashish, I will try working with an erronous record and also a small data load.