Repair FULL to ODS - Cube Data Inconsistent

we have loaded Data into a Historical cube that maintains a " monthwise version " of all data loaded into an Z-R/3 table .The flow is from R/3 table -> ODS - > CUBE . First we did
Initialization and then delta's were running fine for next few months .
one fine day in NOVEMBER, we realised that the data in R/3 and BW was not matching cause
user's were changing the R/3 table from the table maintenance generator and no DATE/TIME
field was getting affected ,so no DELTA was being triggered .
Now we had to correct this NOVEMBER DATA with out affecting the for data older "month versions " which is perfectly alright. This delta data was not available in PSA or RSA7 .
P.S : The " month version " is created at the first load ( i.e on the first of every month ).
If we extract from our infopackage by using selection criteria (by using "november month
version " the complete delta process / earlier monthly versions would be gone( again we need to do reinitialize).
So we changed the load from r/3 to ODS from DAILY DELTA to DAILY FULL REPAIR REQUEST( also
cause NO DELTA MECHANISM WAS GETTING TRIGGERED INCASE OF CHANGES IN TABLE MAINTENNACE) ..i
guess the developer here did a selective deletion from ODS & CUBE before running a "FULL REPAIR REQUEST " .However , the i believe that the activation queue and the ChangeLog remain unchanged during the selective deletion of the data from the ODS , so although the DATA in ODS is CORRECT and has been Updated into the CUBE ..but CUBE DATA IS STILL INCORRECT
Can any of you experts please please help me resolve this issue .
thanks all
anya

Hi,
<i>i believe that the activation queue and the ChangeLog remain unchanged during the selective deletion of the data from the ODS , so although the DATA in ODS is CORRECT and has been Updated into the CUBE ..but CUBE DATA IS STILL INCORRECT</i>
When you do activation after selective upload, the entries will be created in Change log table. So those entries are used as delta . When you are doing Repair request system sends the records as New records.So these records will reach to Change log and Cube as New records only.
So every thing works perfectly,no need to changed the load from r/3 to ODS from DAILY DELTA to DAILY FULL REPAIR REQUEST.
With rgds,
Anil Kumar Sharma .P
Message was edited by: Anil Kumar Sharma

Similar Messages

Error while doing Repair Full Req on ODS

HI ALL,
I am trying to load the data in the ODS starting from 2005 to July 2009 through a Repair Full Req loads. Data source is 2lis_11_vaitm.
I have splitted the loads and wheni try to run the Repair Full Req load into ODS its throwing error in source system.
When i checked inthe R/3 for dumps & Logs it shows
Error when attempting to IMPORT object "MC11VA0ITM_TAB".
"CONNE_IMPORT_WRONG_COMP_TYPE"
"SAPLMCEX " or "LMCEXU06 "
"MCEX_BW_LO_API"
Try and find out why the length of the object is supposed to be
different. There are a couple of possible reasons:
    Perhaps the length of the field imported has changed in the ABAP
    Dictionary. Check to see that the length of the field imported is
    the same as the length of the field in the ABAP Dictionary.
    If the data cannot be reconstructed from another source, if must
    be read and converted using the "old" structure" and then
    re-exported using the new structure so that future IMPORTs
    always work with the new structure.
    A new version of the program is may be active and this may no longer
    suit the dataset. Try to get rid of the error by regenerating
    program "SAPLMCEX ". Proceed as follows: Call transaction SE38 and enter
    program name "SAPLMCEX ". Activate the program with the "Generate/Activate"
    function.
how to rectify this problem.
Thanks and Regards
Janardhan KUmar K
Edited by: Janardhan Karmala on Jul 27, 2009 8:52 PM

There appears to have been a change in your R3 environment, perhaps the application of a Support Package or Support Package Stack, and this change updated the ABAP program for this extractor.
Re-generate/re-activate that ABAP program in SE38. This may require you to get an Access Key from your OSS administrator. Whilst re-generating/re-activating this ABAP program, it may require you to re-generate/re-activate other objects in which case you may be required to get additional Access Keys. A pop-up screen will tell you exactly what objects are needed for the Access Key.

Can we do repair full after re-intiliazing datasources?

Hi Experts,
I work in Production support environment. We are currently involved in bringing back the missed delta records. Currently, we have put our Process chain on hold. We are undertaking Repair Full request activity. Daily, we have been doing manual loading of current data. My doubt here is, since we are doing Repair Full Request, we have not re-initialized.
Can we go ahead and schedule our process chain by re-initializing our datasources? If we do like this, can we do repair full for the past data?
Your suggestion will save our time & work load...............
Thanks In Advance,
Suman Chakravarthy
Muscat(Oman)

Hi...
Are you in BI 7.0......
If yes......it means ...you are executing a Full Repair infopackage till the PSA......you have not touched the delta Infopackage ....right ?
After that you are fetching the data from PSA till the target using either delta or full DTP....right ?
if that the case you don't need to do re-init........delta will work fine......
But the thing is you should not run Delta and Full Reapir in parallel.......technicall you can do that.....
But it will result duplicate records in the cube level......and also if there is any DSO where the aggregation type is Summation......
Once Full repair is done.......then again you can start the delta......
One more point I want to mention here.......if you are not using the Process Chain delta DTPs.......and using any Full DTP......then youy have to run all the delta DTPs which are included in the Process Chain with the option "No fetch;No data".............otherwise the Delta DTPs will again fetched the same records which will result duplicate data......
Regards,
Debjani.......

Using Repair Full to load Data

Hi Guys
I have to load several years of data. Being loaded into ODS and Cubes in delta mode. I want to break the load into small chunks to load in a controlled fashion. That's what I have planned to do
1- Initialize without data transfer
2-Run Repair full for each year from 2001 to 2006
3-Then start running the deltas.
I am just curious if u know if this method is going to work. Does the delta that i m going to run after running the last repair full will bring all the data again or just the changed records.
I also want to go back historically with repair fulls to make the current data available. Is this going to work or not. If not what should I do to load the data periodically.

Hi,
If you want to do this year load to cube then you can full load sepcifying the year in the infopacakge.
You need not run the program which i specified earlier for cube, but in case you want to do this load year full to ODS and in case delta load is going on in the ODS then you will have to first do the full loads for the years u want to and then excute the program and then carry on with delta loads.
I hope I'm clear.
In case you are not able to see the radio buttons then you can drop a mail to [email protected] and tell them that you are not able to give points for this post.
Regards,
Rohini

Duplication of records is possible in ods?? due to repair full request ??

hi,
There was a data load failure for my ods1 to ods2 to cube.
And hence repair full update is been used to rectify it. But my question is that possible while running the repair full update enables to carry out data duplication of records?
Bcos i found in that ods there was duplication of records for 2 days. And how to rectify those duplication of records?
thanks
pooja

Hi,
The chances of duplicate records coming in ODS are less as ODS has overwrite functionality.
Confirm that once.
In case you still feel there are duplicate entries you can perform a selective deletion of data from ODS and then perform a full repair from source system to ODS.
This will make the data correct in ODS.
You may need to correct them in the subsequent targets if they are already loaded.
Follow this blog to correct those.
Repair bad data and subsequent data targets with delta update
/people/aaron.wang3/blog/2007/04/11/repair-bad-data-and-subsequent-data-targets-with-delta-update
Hope this helps.
Thanks,
JituK

Loading through Process Chains 2 Delta Loads and 1 Full Load (ODS to Cube).

Dear All,
I am loading through Process chains with 2 Delta Loads and 1 Full load from ODS to Cube in 3.5. Am in the development process.
My loading process is:
Start - 2 Delta Loads - 1 Full Load - ODS Activation - Delete Index - Further Update - Delete overlapping requests from infocube - Creating Index.
My question is:
When am loading for the first am getting some data and for the next load i should get as Zero as there is no data for the next load but am getting same no of records for the next load. May be it is taking data from full upload, i guess. Please, guide me.
Krishna.

Hi,
The reason you are getting the same no. of records is as you said (Full load), after running the delta you got all the changed records but after those two delta's again you have a full load step which will pick whole of the data all over again.
The reason you are getting same no. of records is:
1> You are running the chain for the first time.
2> You ran this delta ip's for the first time, as such while initializing these deltas you might have choosen "Initialization without data transfer", as such now when you ran these deltas for the first time they picked whole of the data.Running a full load after that will also pick the same no. of records too.
If the two delats you are talking are one after another then is say u got the data because of some changes, since you are loading for a single ods to a cube both your delta and full will pick same "For the first time " during data marting, for they have the same data source(ODS).
Hope fully this will serve your purpose and will be expedite.
Thax & Regards
Vaibhave Sharma
Edited by: Vaibhave Sharma on Sep 3, 2008 10:28 PM

How to do reconcilization of ODS data and CUBE data

Hi All,
How to do reconciliation of ODS data and CUBE data,I know how to do reconciliation of R/3 data and BW data.
Regards.
hari

Hi,
create a multicube based on your ODS and cube, identify commen characteristics and perform the key figure selections; create then a query showing both information from the cube and ODS with perhaps some formula showing differences between key figures.
hope this helps...
Olivier.

Error with calendar date being extracted : PSA - ODS - cubes correction

Hello to all,
I have an issue with incorrect calendar date being captured by BW.
Here is the analysis:
We have red request into the invoice cubes because of incorrect calendar date. It has been found that the process is done via PSA -> ODS -> Cubes
How will I do the correction? Kindly specify on whether or not to change the status of the red QM status in the cubes. Also with the data mart status of ODS and the request of ODS? What status(has to be changed)? what request has to be deleted? Thanks.

Hi,
What I did:
changed QM status to green of the Info Cubes
deleted the green status of requests in the Info Cubes
deleted the Data Mart status of ODS
deleted the request of ODS
edited manually the data into the PSA
started the update immediately
Then....
I will wait if the request will be updated in ODS?
Then
I will check if the delta job from ODS to cube will proceed right?
Please advise. Thanks.

Ods to cube data load verry urgent

Hallo ALl,
i have started the int from ods to cube (data packet 1200000) and its talking hell of time there are 85 data packet, can some suggest me how can i seep up the data load.
regards sanjeev

hi
1) Stop the load first.check the job log and st22 if htere any shor dumps
2) go to the *8 gnerated infosource...
3) createa New IP.with the option as PSA and datatarget
4) it will take little time but it wont fail..
5) this will process package by package...
hope it helps
regards
AK

Diff between init with data transfer and repair full request

hi,
i have observed that even in the new flow we are doing init without data transfer and then repair full request
if i do init with data transfer also i can achieve the same?
i want to know why we need to do this ,do we have any advantage of doing init without transfer and repair full request?
please suggest me

Hi Venkat,
A repair full request is loaded in cases where you get erroneous records or where there are missing records which you want to load. In such cases, a repair full request for the selective records can be loaded and is the most efficient way to correct the data, since you won't be required to load the entire data once again. Also you acheive the desired result without disturbing the delta. (You are not required to do an init w/o data transfer after a repair full, people just do it a a precaution)
However, repair full requests should only be loaded to infoproviders with update mode 'overwrite'. Otherwise, if the InfoProvider, its very probable that you might double the value of the key-figures due to rows being added twice - in case of an InfoProvider with update mode 'Additive'. So, if your InfoProvider is additive, you will need to delete the entire data and do an 'init with data transfer' to avoid corrupting the data. (Note: you can do a repair full request for an additive infoprovider in case of lost records or if you can delete erroneous records with selective deletion.But you have to be careful with the selections lest you inadvertently load other records than required and corrupt the data)

How to correct Inconsistent Cube Data

We have following unique condition of the info Cube data which is causing duplicate records in reporting.
The Fact table of the Cube has data for one dimension ID 8042 in the Data Packet dimension.
This dimension Id does not appear in the Data Packet Dimension Table. But still the data is available in the fact table which is supposed to be a deleted data.
How can we delete this wrong dimension ID data from the fact table without deleting the whole Cube contents? We can not use selective deletion because we can not select the DIM ID in it and the 0REQUID for that dimension does not exist.

Mangesh,
This could be caused due to "automatic request deletion" option in the InfoPackage while loading new data.
You can use a function module RSCDS_DEL_OLD_REQUESTS to delete such requests from specific InfoCube.
If you are using Oracle database, you can use program SAP_DROP_EMPTY_FPARTITIONS. It deletes partitions (requests) that are no longer referred to in P-dimension.
Hope it works...

How to delete repair full request

Hi,
There was a delta load scheduled in background. By mistake before the delta load a repair full load with particular company code selection was run. So after the delta load, data is duplicated in the system. Now the request is compressed. I want to delete this repair full load request from the ODS. Please let me know step by step solution for this.
Thanks.

Hi,
from where u r loading data into cube from ODS or R/3. if u run a repair full load then it will not check the data mart symbol although it is successfull so the next delta will brings data again, so u have to run init with out data transfer to make deltas 0 in source. this happened in ur case .
but the cube is compressed so u cant delete data request based
is there any PSA there if there is PSA then u can use REVERSE POSTING from PSA this will brings key figure values with - values so the records will be nullified
in manage of the cube -> select that repair full request -> click on monitor button -> u can find an icon in the top(psa request symbol with red mark) click on that and execute.
if there is no PSA then u have to do selective deletion.
assign points if it helps,
thanks,
pavan.

Add new key figs or characteristics and "Repair full request"

Hi Gurus, here are my questions
1.If I add new key figs or characteristics(which are not Key fields) to ODS and Cube containing tons of data, Will the transport for structure changes fail due to presence of data?
2.The extractor to the ODS which inturn is feeding the cube, is delta enabled. So the newly added objects wont have entries for historical records. If I change this extractor to "Repair full request" after adding the new fields, will the new fields then get populated?
3. After the "Repair Full Request" activity, Can I change the Infopackage setting back to delta update mode?
Thanks for the help in advance!!
Simmi

Hello Simmi,
1. Transport should go fine. As you are not changing the key fields, it should not cause any problem even if data is present.
2. Yes. But make sure what data you want to laod. As with full repair, system is not going to check for duplicate data. The responsibility lies with you.
3. Sure. if you mark the full load as repair, the delta settings will be unaffected.
Hope it helps.
Regards,
Praveen

Diff b/n Repair full and Full load

Experts
Please explain diff b/n Repair full request and Full load.
If I do Repair full request instead of Full load ,is there any problem in the data ?
Many thanks
Manoj

Hi Manoj,
You can use a request that was selected as a repair request via Scheduler ® Repair Full Request to carry out a full update in any data target. This also applies to data requests that already contain data from an initialization run or deltas for this DataSource/source system combination, and that have overlapping selection criteria.
The difference between a Full Load and a Full Repair load is only evident in the case of loading to an ODS (to which you are also loading Init and Delta from the same datasource). If you ODS has the Init request active and you load a Full request, the data will not be activated. You should always use a Full Repair request when you want to load a full load to the ODS. For a cube it does not make a difference as the sequence of data loads is in important in a cube.
If you have accidentally not marked the request as a repair request, you can convert is using SE38 > RSSM_SET_REPAIR_FULL_FLAG
Hope this helps...

Repair full load

Hi,
I have "Billing document" number( 9000182) in ODS .
For the debugging purpose I want to do " REPAIR FULL LOAD' only for this document .
can i do repiar FULL LOAD" for this documetn in Production system?. Is there any problem if i do repair full load. any data mismatch occurs.
I can do selective deletion and i can do " REPAIR FULL LOAD' but my doubt is any problem will come as it is ' PRODUCTION System".
Please confirm .U r suggestions will be helpful.
i tried in Q system but that document was not there in R/3.

Hi Kotha,
If you are doing this for debugging purpose,Then no need to delete the data from ODS.
Follow the below steps.
1) Run IP with options "load till PSA" and selections for Billing no 9000182(If you think it will run for longtime,do full repair or run it as full load)
2)If you want to check thi data in CUBE's as well then run in simulation mode using Update rules with respect to each target.
Hope this helps.
Regards,
Venkatesh

Repair FULL to ODS - Cube Data Inconsistent

Similar Messages

Maybe you are looking for