on 2014 Feb 25 2:35 PM
Hi All,
I am loading data to Datasource from three infopackges with three selections.(say fiscper 2012, 2013, 2014)
Through DTP i am laoding data from Datasource to INFOCUBE. Then three requests will be there in cube.
While for next laod , I have to delete the overplapping request from cube for that paticular selections
(means the request updated in cube for ficper 2012 selection in infopackage must be deleted and the latest request for fiscper 2012 must be in cube, likewise for 2013, 2014)
Only three rquests must be there.
Can u help me how to do it?
Thanks,
Karthik irkamsetty
Request clarification before answering.
Please check this blog for solution you are looking for
Using Routine in "Delete overlapping request" vari... - SAP Community
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Unique requirement, but I am agree with the suggestion given by Arun Thangaraj.
I will suggest :
1. Create 3 DTP with same selection you have for 3 IPs which are used in PCs
2. Those 3 DTP must be FULL, otherwise the delete overlapping request will not work
3. Since it is FULL DTP you have to make sure deleting the PSA / Take help from ABAPer that you want to load data for that run date for daily PC/week for weekly PC and so on
Thank-You.
Regards,
VB
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Karthik
Whether the infopackage selection is hard coded or through a routine?
Try using selections in the DTP also so that you can always delete data based on the same selection, this can be achieved through a DTP routine or three different DTP's with different selection.
Note: If you have the same selection in DTP then system will delete only the last request and not other requests.
Regards,
Sakthi.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Karthik,
So, from what I understood, you have 3 different process chains loading to the same cube. So, in the first chain, after info package triggering step and just before the DTP trigger step, add one more step for "deletion of overlapping request". This process is available in RSPC when you open the chain in change mode, and just give the required parameters and ask it to delete all the old requests. DONT ADD THIS STEP IN ALL THE 3 PROCESS CHAINS. Just add this step in the first chain only. So, everyday, after running of these process chains, you will be having only 3 requests which will be showing the current data.
Please let me know the feedback and dont forget to assign points if this is helpful to you
Thanks and have a nice day
RJ Renju
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi
I would suggest to keep three FULL DTPs as similar to info packages and keep the same selections in DTP as well from info package. you have to load these three DTPs one by one to CUBE.
Now you have 3 requests in cube as well with selections. This way you won't miss any records from PSA, also you would be able to do overlapping request from cube with same selections to keep only last three requests. Hope this helps you.
I assume you would delete PSA everyday before you do data loads for next day
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
If you don't delete the PSA data, then add logic in start routine to delete the duplicate data; always latest request's record comes in the last...so you can compare based on the keys and do a deletion to keep the latest records,
Say you have loaded 3 requests with year 2012..so you maight have three records for same key in PSA....
and in start routine you can add logic to delete the 2 records updated by old request ....
or you can have PSA -> Cube DTP as "Delta request one time Only"..so that you won't duplicate the values in cube, also no need to delete the PSA
Hope this gives an way
Hi karthik,
Have you create the PC for these cube? please let me know.
if you created the PC on cube you can use delete overlapping request it will delete total data then load based time selection.
Thanks,
Phani.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Karthik,
Your loading the same request multiple times to cube.
while PSA to cube make a filter fisper 2012, 2013, 2014.
cube is addition property means delete the overlap request then load.
Thanks,
Phani.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Karthik,
You may model this in process chains.
After you load the 3 new new requests, use the process step delete overlapping requests and give the parameters that you require.
Let me know if you need more details on this.
Shilpa
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi shilpa,
Cube is laoding through three different process chain.
One process chain is laoding daily, another is weekly and another is monthly.
All are loading from same datasource through DTP(without selections in DTP) but process chain has three differnt infopackages with different selections.
While for next laod , I have to delete the overplapping request from cube for that paticular selections
(means the request updated in cube for ficper 2012 selection in infopackage must be deleted and the latest request for fiscper 2012 must be in cube, likewise for 2013, 2014)
Only three rquests must be there.
can u help me how to do it. I am new to SAP BI.
thanks
karthik
Hi Karthik,
If the process is handeled is 3 different process chains, you can write a routine to manage the deletion criteria, you may take help of ABAPer to help you write the code.
Mantain the selections as shown in the screenshot
Here you can select the deletion criteria, based on your needs.
Thanks
Shilpa
Hi,
Every time your going to load 3 request to with proper selections. in that case why can't go with avaibale 3 request deletion and load 3 requests with proper selections. it will be same as you need.
check with overlapping request deletion from cube options(conditions and exceptions) at process chain level.
Thanks
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Not clear, you have 3 data sources? or only one data source with 3 info pack(different selections).
your cube daily having 3 request only. am i right.
if yes then just delete the overlapping requests based on data source or source system. ant any one of your process chain just add process step - delete overlapping request from cube.
do test case at dev system.
please search on Google and read the document which explores the same.
use search term "delete overlapping request from cube".
Thanks
Cube is laoding through three different process chain.
One process chain is laoding daily, another is weekly and another is monthly.
All are loading from same datasource through DTP(without selections in DTP) but process chain has three differnt infopackages with different selections.
While for next laod , I have to delete the overplapping request from cube for that paticular selections
(means the request updated in cube for ficper 2012 selection in infopackage must be deleted and the latest request for fiscper 2012 must be in cube, likewise for 2013, 2014)
Only three rquests must be there.
can u help me how to do it. I am new to SAP BI.
thanks
karthik
Hi karthik,
How about make some dummy DTP for difference chain? (i.e. setting is the same, with difference name only)
And make use of the "Only Delete Request from same DTP" option?
It should be available under Deletion Selections, if you use DTP as object type for Delete overlapping Request from InfoCube.
Regards
Bill
User | Count |
---|---|
72 | |
18 | |
10 | |
7 | |
7 | |
4 | |
4 | |
4 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.