on 04-16-2013 9:42 PM
Which Error Handling option is better based on you experiences and why? I am looking for a Best Practice.
Loading Data from DataSource to DSO
1: No update, No reporting
2: Update Valid records, No reporting (Request Red)
3: Update Valid Records, Reporting Possible (Request Green)
Loading Data from DSO to Cube
1: No update, No reporting
2: Update Valid records, No reporting (Request Red)
3: Update Valid Records, Reporting Possible (Request Green)
Appreciate you feedback.
BW
Hi Team,
From Data source to DSO Use.
Update Valid records, No reporting (Request Red)
Why - all correct records move to target and erroneous records will move to error stack. reporting won't show. bcaz whole records are not updated. some records(error records in error stack). thru error dtp move those records to dso once the correction was done and make first request as green.
Loading from DSO to Cube
You can use option 2 or 3. 3rd will be the best one. once you know that we won't face any error records from dso to cube then 3 rd will be best one. once the load was success, data will available for reporting.
most of the error records will rectify at before loading dso or before activation dso request.
so no need to use error stack again in this data flow.
Thanks
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Team,
if you the get the above error. then you can't load data. need to solve that issue then only it move to prod or dso activation will success.
Please go thru be lowl inks. its give you clear settings.
Thanks
Hi,
From my past experience, i am sharing you this. In my project, the users needed data whichever is available at the moment. For Instance, for a particular plant/store, if there are 100 records added per day, out of which 50 are stuck in error stack(Datasource to DSO), they need the other valid 50 records in their reporting. So, we had the option, Update Valid Records, Reporting Possible (Request Green).
If in case, the users needed reporting in such a way that, even if there is one erroneous record for a particular store/plant combination(or your semantic group selection), they dont want any records to appear in reporting for that store for that day. So, in this scenario, we can use No update, No reporting. Then
So it depends upon the user's requirement.
Also, as Raman told, be sure enough that all error records are captured at datasource to DSO level, so that you can use option 3 for DSO to cube.
Regards,
Pratish
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
86 | |
10 | |
10 | |
9 | |
6 | |
6 | |
6 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.