cancel
Showing results for 
Search instead for 
Did you mean: 

Selective deletion and repair full request!!!

Former Member
0 Kudos

Hi,

I have one DSO which is receiving deltas from Std extractor and is also serving deltas to many data targets (Both DSO and Infocubes)

Now there comes the need of deleting data in DSO by means of year ( selective deletion )

I know that i need to adjust the data flow from DSO to other data targets myself!!

So after doing selective deletion in DSO, can i update datatargets ( of 3.5 version ) with full update so that data is adjusted in correct format for the subsequnet data targets?

or is there any better option??

Accepted Solutions (0)

Answers (5)

Answers (5)

former_member182869
Participant
0 Kudos

Dear Simran,

You have to do selective deletion to all subsequent Data Targets manually only or you can there is no other alternative in that or you can drop the data and reload only Current Data which is there in Original DSO

from next loads on you will get only deltas regularly,

But one thing i am unable to understand that why single DSO serving same data to different Data Targets, Is there any authorization issues ? Brief me .

Regards

Sree

Former Member
0 Kudos

Hi Sriram

Dont know why you want to do a selevtive deleltion.Is it due to the wrong or duplicate data?

After doing the selective deletion due to to load the data for that period from sourcextractors.

Once you had done with the selective deletion on DSO with respect to 0fiscper,then u need to do selective deletion for the above targets aswell for the same periods.

And then if daily deltas need to continue its always a good practise that you do a Init without datatransfer from this DSO to further targets.( But before doing this Init be sure that all the delta's are updated further)

I guess this should help out

Regards

Jagadish

Former Member
0 Kudos

Hello,

can any one explain why we should use repair full request, we can use FULL update also right.

can any one justify briefly.

Thanks in advance.

Regards,

Sujith

RamanKorrapati
Active Contributor
0 Kudos

Hi Sujith,

Repair full load - Just brings missed delta records from source.

full load - Its brings whole historical data which lied in source system.

Even you can go thru search on scn and google it. you will get lot of info.

Thanks

Former Member
0 Kudos

Hi Sujith,

Difference between Repair Full and  Full Load ;

1. When we want to pull the history data from the source system(SAP R/3) for the first time till yesterday and trigger / enable the Delta load.

For Eg; for your client,when the implantation was done ,when the GO Live started, all the history data(sales,Material,Finance etc) would have pulled from the source systems into BW system and enabled the delta load from next day.

Note: In this scenario ,We use Full Load only for the one time.

2. In most of the scenarios we will use Full Load for the master data (Daily Load). In few scenarios we will use Full Load for the transaction Data Loads. For Eg: in case if there are no changes happening in transaction data daily (like sales data) ,then we go for Full Load for transaction Data Load also.

3. If someone has deleted the data in the data target by mistake which was loaded using 'Full Load' ,then we can go for re-loading the lost data using the same normal 'Full Load' option. In this case if you are using a DSO in between PSA and Infocube then if you don't use the selection parameters(Company Code,Profit Center, Sales Organization etc) for the lost data ,then also it's fine. Because DSO has overwrite option. In case ,if you don't have DSO,then you have to use the selection parameters on infopackage/DTP level depending on from which level you are trying to get the data back. i.e Source System or PSA.

4. We have scenario where the data which was loaded using DELTA load has been deleted from the data target, OR due to DELTA load failure the data didn't get loaded into the data target. Now,we want to get the lost DELTA data back into the target. In this case we can go for using both 'Repair Full' and also 'Full Load'. but depends on the below conditions;

If we want to use 'Full Load' to get the lost DELTA data,then we should block the users in source system(SAP R/3) from posting the data,and clean up the delta queue . Then stopping the daily process chain in BW system which will bring the DELTA data. Then we can use' Full Load' option.

But, If we don't want to disturb the daily DELTA load and if we don't want to block the users in source system, then we should use 'Repiar Full Request". Using this option, the daily DELTA load will not get affected.

Please let me know you are clear with my points.

Regards,

Asha

Former Member
0 Kudos

Thanks ramana

Former Member
0 Kudos

Hello Asha,

Thanks for details information.

yes, you have answered 30% of my question.

i found the actual answer in some other blogs.

the main reason for full repair is

  1. it wont disturb existing delta
  2. when we load with full update the dso activation will not happen due to current loading will delta, so in these case to get missing delta we should pull data with full repair mechanism.

Regards,

Sujith Sadige

Former Member
0 Kudos

Hi........

If u do Selective deletion from the DSO............if u hav to delete the same records from the Furthur target also.........to save data inconsistency............Full update will not help.........by full update u cannot delete that request.......

Moreover if u do full update in a cube............there will be duplicate records..........cube is always additive.........

In case of ODS if the update mode is Overwritten.............then u can do..............but it will not help...........moreover DSO does'nt support Delta and Full upload parallely.............DSO activation will fail.........In that case u hav to do Full repair.......

Regards,

Debjani.........

Former Member
0 Kudos

OK here is more explanation:

I have one DSO which is having millions of records nearing 100

Which is now receiving deltas from Std extractor ) and also serving deltas to many data targets (both DSO and cubes )

Becasue of huge volume of data we have decided that we will only maintain the data which is current year and previous year and rest of the years data is moved to some other DSO ( copy of the original one )

After moving the historical data to the new DSO we should delete the historical data

Now after deleting the seletive data , how should i adjust the data in subsequent data targets??

as the subsequent data targets are nearly 30 in number it is a daunting task to go to each and every target and do the selective deletion

What i am thinking is

1) do a selective deletion in original DSO

2) creata a copy of the DSO and update data other than current year and previous year

3) Selective deletion from the Original DSO

4) Do the same sective deletion for the connected data targets

5) Update the data targets with full upload ( repair full request )

the point 4 which is making me worry as it has too many data targets

is there any better work around for this?

please suggest me

Former Member
0 Kudos

Hello,

Still not clear what you want...but just rememebr then cube will add and not overwrite...so whatever you load to the cube as full load should be deleted first from the cube through selective deletion and then loaded or the records will get doubled.

If its just a selective deletion from the DSO and no new loads to the DSO then you may have to be careful as delta for those records may not go correctly for the targets

Thanks

Ajeet

former_member207028
Contributor
0 Kudos

Hi Simran,

So after doing selective deletion in DSO, can i update datatargets ( of 3.5 version ) with full update so that data is adjusted in correct format for the subsequnet data targets?

you need to delete the same data in Cubes also with selective deletion, in caase of DSO not need because of overwrite functionality, but better do it in all data targets.

the with full update you can load to DSO and all data targets

Regards

Daya Sagar