Delta load from ODS to Cube

[Former Member]

Hi,

Daily we are doing delta load from ods to cube.

Accidentaly I have deleted todays delta request from the cube,

to get the same data again into the cube I did following steps.

1.In ods the request which is updated into the cube made it status not okay again updated data from PSA to ODS.

2.When I tried to delta load manually from ODS to cube,

I'm getting error saying that "Error in source system" and "Delta update ODS is invalidated".

whetherI need to delta initialize again into the cube?..

Rgds

shankar

[edwin_harpino · Accepted solution]

hi Shankar,

welcome to SDN ...

take a look

hope this helps.

[Former Member · Accepted solution]

Hi AHP,

Thanks alot for your information,

I have some doubts to do re initialize again.

I need to delete previous init pointer from the cube before doing reinitialization? and also what about old delta request.

Rgds

shankar

[Former Member · Accepted solution]

Hi Siva,

U could have simply reconstructed the request in the cube.I mean the deleted delta request.

Here as u have deleted a request from the ODS,which is giving deltas to the cube,the 'delta maintenance' sequence that was maintained between ODS and cube has gone out of sync.

Now suppose that delta load from ODS to Cube was wrong and u wanted to delete error request from cube,and reload it from ODS.

So what u do is,make request RED in cube,then delete it from cube,and then whichever requests were loaded in that delta from ODS to Cube,for those requests..click on the tick mark that appears near the request number,a new screen will appear,there click on the delete icon.It will give message 'reset delta maintenance for this request?' confirm as yes..refresh in manage screen.tick mark will disappear for those requests.DO NOT DELETE THE REQUESTS.Now run a delta load from ODS to Cube.Load will go ahead and pick delta from required requests.

Now in ur case..u deleted request from ODS without deleting the tick mark..

Go to the infopackage to load init from ODS to cube..go to menus..Initialization Options from Source System and delete the entry for init that appears there..

Now go ahead and run the init infopackage with settings of 'Init' and check the checkbox for 'without data transfer'.It will send a 1 record request from ODS to cube.

Now reconstruct the request u had deleted from ODS.Tick mark will not appear for it as it isnt updated to data target.

Now run delta from ODS to Cube.Delta will pick up the delta from ODS and tick mark will appear for the request in ODS.

Sunsequent deltas will run OK..

No need to delete previous init request from cube.

Old delta requests will be all OK..

cheers..

Vishvesh

[Former Member · Accepted solution]

Missed one point..

Before running init package also delete today's delta request from the cube.

cheers,

Vishvesh

[Former Member]

Hi Shankar,

As far as I understood, you make a request in the ODS not OK. Did you delete it?

If not, the subsequent request from the PSA will not be activated. You need to delete the red request from the ODS and activate the last request. After that try to load data into a cube again.

Best regards,

Eugene

[Former Member]

Hi Siva,

If you are loading to Cube from ODS via PSA then

you can just do a reconstruction in the infocube for the

delta request you have deleted from the infocube. As per

your explanation it is clear that you have made the

todays delta request in ODS red from green. If possible

make it again green and load the delta from ODS to Cube.

Regards,

Prakash B

[Former Member]

Prakash,

While loading from ODS to cube we will not have PSA option.

Regards

siva

[Former Member]

Hi Shankar,

Why did u made the Delta Request of ODS as not okay if it is loaded properly to ODS.

From your message i cld see delta load to cube has been deleted. If you have just deleted that request, then you can Reconstuct the same request which you can see in Reconstruct tab.

Before that if teh data load to ODS is find then it to OKay first.

Hope this helps. i hope i have understood ur error clearly.

Cheers!

Bhuvana.