cancel
Showing results for 
Search instead for 
Did you mean: 

data extraction via combination of full + init updates

Former Member
0 Kudos

in order to reduce performance problems we're currently experiencing, as well as to spread out the resource drain of data extraction, i'm planning to perform full updates for infoproviders on a per month basis (based on document date), and then perform an init delta on the last month.

what possible problems might i encounter if i setup my processing this way, besides not being able to track any changes that were processed in full update?

thanks.

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

If I understand you right you want to do the initial upload by doing a delta init based on document date for the last month and upload the previous months in full upload. The biggest problem you might have is that you may not get changes on older line items, e.g. cancellation of old documents that often doesn't change the document date. If this cannot happen at your installation you could go with it. An alternative is to do an initial upload without data transfer for the last two years and then do a full upload on a monthly basis including the current month. You need to make sure that noone is able to post in the system during this procedure.

Best regards

Dirk

Former Member
0 Kudos

yes, that's right. in that case i think i may get by with the approach i suggested. i would like to understand the approach you are suggesting though - if i did that, would past data (which we loaded full) also be taken into consideration with delta updates? just to be clear:

1- i have october and november data.

2- i init delta w/ no data transfer.

3- i full upload october and november.

if i made any changes to october or november, would a delta transfer capture these changes?

klaus_werner
Active Participant
0 Kudos

Yes it would.

Regards, Klaus

Former Member
0 Kudos

i just tried this today, and found out that i could not do this for an ODS - i was trying to update an ODS with three different extractors, which, when activated should consolidate lines based on their key. in my case, i'm using 2LIS_02_ITM for PO items, 2LIS_02_SGR for GR quantities, and 2LIS_02_SCL for the delivery date. during delta loading, a PO item would have one entry for each extractor, and upon activation, consolidates into one line (since the data extracted from these three extractors don't overlap, save for the key fields). it refuses to work when i load via full update.

is this what is really to be expected?

Former Member
0 Kudos

I think I had the same problem once. Are you unable to upload Full and delta updates into the same ODS? In this case you might want to create an identical ODS only for the full uploads.

IIRC you are able to start with the full upload and then do the delta init but that's not what you want to do.

Best regards

Dirk

Former Member
0 Kudos

darn. there goes that brilliant idea. 😃 i was hoping to be able to perform an init w/o data transfer for the ODS on PRD, perform a flash copy of PRD, extract the historical data from flash copy (via full uploads), and once that's complete, return to PRD for delta updates. but since it doesn't look like it'll work the way i was hoping it would, i guess i'd have to look some other approach...

but would this same effect occur with infocubes? as in, not be able to load the in the way i was hoping to load data?

Former Member
0 Kudos

I think this works fine with InfoCubes. As far as I know there are only restrictions on the ODS because you could mix up your data with all the full and delta uploads if you write it into a cube.

Best regards

Dirk

Former Member
0 Kudos

hi dirk,

just thought you'd want to know - that i was able to successfully integrate full updates (from historical data)and delta updates (present data) into the same ODS.

i was able to do it by implementing OSS note 689964 (it's a program actually, that makes the full update requests of an ODS look like repair requests), activated them, and successfully consolidated everything into its cube.

of course i have to make certain that i don't make the silly mistake of full loading present data, but the setup table maintenance helps prevent that from happening easily.

hope this would help anyone else who might encounter the same problem. 😃

ryan.

Former Member
0 Kudos

Hi Ryan,

thanks for the information. It's really helpful. Good that you were able to do the upload.

Best regards

Dirk

Answers (1)

Answers (1)

Former Member
0 Kudos

thanks a lot - i'll keep your suggestion/s in mind. 😃