Need to delete data...? for enhancing datasource..?
Sub : Need to delete data...? for enhancing datasource..?
No need to delete any data in BW or R/3 side for enhancing datasource if loading data into ODS in overwrite mode.
Simply follow the below steps:
1. Add new fields to ODS & cubes and adjust update rules.
2. Clear LBWQ(update queue - Run V3 job only).
3. Clear RSA7(Delta queue - Run infopak to pull data into bw)
4. Move datasource changes to Production, replicate and activate transfer rules.
5. Delete data in LBWG(setup tables).
6. Fill setup tables for historic data.
7. Initialize datasource if required, without data transfer(Zero initialization).
8. Pull data from R/3 to BW ODS in overwrite mode with Repair Full option.
Loading data in overwrite mode, so no problem, Just load again historic data as well. and push delta from ODS to further(ods/cube).
9. Push delta from ODS to CUBE.
More info @ https://www.sdn.sap.com/irj/sdn/thread?threadID=624282&messageID=4393416#4393416
5 Responses to Need to delete data...? for enhancing datasource..?
hi,
Nice blog..
I have a question. Tell me one thing this steps you're telling is BI7.0?. Because as you mentioned in Step 1
"Add new fields to ODS & cubes and adjust update rules" .
We cant add new fields in the Cube or ODS unless we delete the data in 3.5
Satish,
In 3.0 or 3.5 also to add Datafields to ODS or fields to cube in same dimention(not new dimension) we no need to delete data. We can add directly. I done many times.
Give a try in Development system and update me...!!!
Hi Srinivas,
Since all the data is to re loaded again, irrespective why not delete data in ODS prior to loading new enhanced data? ( which can improve the load performance by not checking each new data record with existing data in ODS, which is in overwrite mode).
One more question, if the data is further updated to cube from ODS , first cube data should be deleted rite before loading new data from ODS??
Very nice Blog.. Really appreciate your efforts
thanks
Rajesh
Dear Rajesh,
Different approches for different requirements/cases. For every change request implimentation, deleting data and reloading is not correct approch(data will not be available fore reporting).
For improving load performance we can delete data(to avoide huge change log). In this we need to delete data from CUBE aswell and data will not be available for reporting. I dont think it is a good approch.
2nd Question : No need to delete data from CUBE. Before loading into an ODS initialize W/O data transer and loade ODS and push delta to cube.
Hope it Helps
Srinivas Neelam
hi neelam , my issue is like this
I have a problem ,the scenario is like that
we have top level : cube 1 having 4 data sources one of which is cube 2
2nd level : cube 2 and having only one DSO as data source
3 level : dso 1 as having 3 data sources one is a load from itself.
so the flow is like that dso1->cube 2->cube 1
now the issue is i have to update the data all through this model without disturbing the existing delta.
from dso1-> dso1 i can have repair full request .
please suggest the way to update other cubes.
Thanks,
Sapsaps
Something to say?