SPM - Help needed for BW datasource loading option

[Former Member]

Hello,

We are trying to prototype a solution and have hit a bottleneck in terms of loading data through BW datasource option.Need some help on the following doubts:

- Datasources are created in the ECC side, replicated in the BI side, relevant entries are maintained in the OPM_Sources tablej, webservices are all active and RSA3 run of the same extractors in the ECC side shows data. SPM UI perform dataload option does show up option of loading data as per the entries maintained in the OPM_SOURCES , however, it results in error ( Validation - Error , Loading - Error ) the moment load starts. Any clues ? also from where can we learn more about the errors and the causes ?

- OPM_SOURCES - Nomenclature followed is as below:

ID - Tech name of source system##datasource name

Source Name - Object_DS - *Any special nomenclature to be followed here

?*Type - As defined in OPDM_Type table

- While loading data into SPM through BW datasource option, is it still mandatory to define import and export folders in the global settings ?

- Global settings - While defining the source system in global settings , does one need to define both the ECC as well as the BI system ?

Any advice on the above points will help us proceed and know the product bettet.

[Former Member]

Hi Gundeep,

To register a datasource in OPM_SOURCES table you have to populate 4 columns, also documented in OSS note 1239883

SOURCE_ID = <Physical Source System>##<DataSource Technical ID>

SOURCE_NAME = a technical name for the data source

UPLOAD_TYPE = the upload type that this DataSource updates (INVOICE, PO, CONTRACT, ITEM, etc...)

SOURCE_TYPE = "DS" (DataSource)

Below is an example:

SOURCE_ID: ERPCLNT090##ZINVOICE

SOURCE_NAME: INVOICE_DS

UPLOAD_TYPE: INVOICE

SOURCE_TYPE: DS

Yes Import / export folders are necessary. Also you need to have at least one 'System' created and activated under global settings. Similarly have one upload created and activated.

Against every datasource registered in OPM_SOURCES table on the BW side make sure you have one and only one info-package.

When you trigger a load from the UI if the system indicates errors in the UI, check the status of the job that was create on the BW ABAP side, that should give you a good idea where its failing.

The system first triggers the info-package to get the data from your source to PSA, then use API's to load form PSA to the target.

Hope this helps.

Rohit

[Former Member]

Hi Rohit & Rajesh,

I am working with Gundeep on the same assignment.

I had few question on the same lines.

I have my ECC source system ready with the flat files generated for different objects. I have my BI system ready with data source and corresponding cubes related to SPM installed. I have my SPM system up and connected to the BI system.

My question remains at what point do we get the entries reflected in the OPM_SOURCES or do we have to manually maintain it, if manually then, this table does not have active maiantainance wizard (sm30). We realize that this table is very important as entries from this tables are reflected in SPM system and base for SPM data loading.

Regards,

Sampat Desai

[Former Member]

Hi Sampat,

There are 2 ways of loading data into SPM:

- Using flat files

- Data sources

If you plan to load data using flat files generated in ECC, you need to place them in the import folder specified under the 'Global Settings' general tab.

For loading data using data sources you need to maintain entries of the SPM data sources in the "OPM_SOURCES" table. For more details on how to register data sources please refer to OSS note 1239883. Yes you need to manually maintain entries in this table, if a table maintenance screen has not been provided please create one.

Regards,

Rohit

[Former Member]

Hi Sampat

After you have replicated the data sources on BW side, you can maintain the entries in the OPM_SOURCES table.

You could either go to SE12 and enter the values directly or if system restrictions dont permit you to do so, then create a simple z program to make those entries for you. Remember, you just need to maintain the entries once per data source. No customizing view is provided for this table.

The above explanation is only applicable for data sources and not flat files. If you have flat files, just place them in the configured import folder for SPM and they will show up in the SPM UI, and then you can take it from there.

Regards

Rajesh

[Former Member]

Good thread here but a general question for direct update datasources from ECC, can we not use the SPM UI, thus turn the direct update DSO's into LSA methodology and then link directly to the ECC datasources, thus following SAP Best practice?

[Former Member]

Hi,

Yes, you can change the inbound layer direct update DSO objects to regular DSO objects and connect the ECC data sources to load data using BI mechanism. But there are a few things that you have to change to use them in conjunction with SPM data management.

Regards,

Rohit

[Former Member]

Hi Rohit,

Thanks for the reply. We are trying to take the emphasis away from the SPM Data management tool and thus use the BW mechanisms themselves, that way automate the loads via bw and not via the ui in spm. By a few things to change, can you elloborate on these? Any entries to tables (both on source and BW) or just add trfn's , dtp's, datasources, infopackages, etc.

Thanks,

Pom

[Former Member]

Hi Rohit

I have similar questions to Pom in regards to changing the inbound layer and perhaps to move away from SPM Data Management Tool, for the direct update method. I believe the management tool is there to accommodate direct update and flat-file loads from one place, which has its merits. One option we have been looking at is to implement the LSA approach to the existing SPM content model and to have the loads from ECC managed using existing tools, not the SPM UI.

We may have a requirement to load data from non-sap sources at a later stage (using the SPM UI), so would I be right in thinking that we should leave the direct update DSOs in place to allow flat-file loading via the UI and build a seperate inbound layer to manage the direct update loads if we're going to create our own tranformations from the content datasources?

Also on the topic of datasources, how are the deltas managed using the Tool? I have noticed that the transactional and several of the large master data sources have a date selection available, so would therefore assume deltas are managed via this? What would be the best practice for managing these if we chose the LSA approach? Creation of a dynamic selection within the infopackage which could align (or slightly overlap) with the frequency of extraction for example?

Many Thanks

Andrew

[Former Member]

Hi Gundeep

1) The correct configuration for BW DataSources in SPM is explained in detail in the documentation in the Extractor Starter Kit note 1239883 (Direct_Update.doc, chapter 1.3). As long as you have followed the instructions, it should be fine.

2) Did you create the infopackage?

3) SOURCE_NAME: This can be anything, thats appropriate to you.

4) Does the extractor error out immediately or after a little while? Does the data come to the PSA?

5) Yes pls do create the folders, this is a certainly a prereq before you start loading data

6) Source system represents the conceptual sources of data, in which case BW is obviously not

Regards

Rajesh