Tuesday, February 17, 2009


Object documentation DataSource
Locate the document in its SAP Library structure


Set of fields that provide the data for a business unit for data transfer into BI. From a technical viewpoint, the DataSource is a set of logically-related fields that are provided to transfer data into BI in a flat structure (the extraction structure), or in multiple flat structures (for hierarchies).

There are four types of DataSource:

DataSource for transaction data

DataSource for master data

DataSource for attributes

DataSource for texts

DataSource for hierarchies


DataSources supply the metadata description of source data. They are used for extracting data from a source system and for transferring data into BI.

The following image illustrates the role of the DataSource in the BI data flow:

This graphic is explained in the accompanying text

The data is loaded into BI from any source in the DataSource structure. An InfoPackage is used for this purpose. You determine the target into which data from the DataSource is to be updated during the transformation. You also assign DataSource fields to target object InfoObjects in BI.

Scope of DataSource Versus 3.x DataSource

3.x DataSource

DataSources were previously known as DataSource replicates (object type R3TR ISFS) in the BI system. The transfer of data from this type of DataSource (referred to as 3.x DataSources below) is only possible if the 3.x DataSource is assigned to a 3.x InfoSource and the fields of the 3.x DataSource are assigned to 3.x InfoSource InfoObjects in transfer structure maintenance. A PSA table is generated when the 3.x transfer rules are activated, thus activating the 3.x transfer structure. Data can be loaded into this PSA table.

If your dataflow is modeled using objects that are based on the old concept (3.x InfoSource, 3.x transfer rules, 3.x update rules) and the process design is built on these objects, you can continue to work with 3.x DataSources when transferring data into BI from a source system.


As of SAP NetWeaver 2004s, a new object concept is available for DataSources. It is used in conjunction with the changed objects concepts in data flow and process design (transformation, InfoPackage for loading to the PSA, data transfer process for data distribution within BI). The object type for a DataSource in the new concept - called DataSource in the following - is R3TR RSDS.

In DataSource maintenance in BI, you specify which DataSource fields contain the decision-relevant information for a business process and should therefore be transferred. When you activate the DataSource, the system generates a PSA table in the entry layer of BI. You can then load data into the PSA. You use an InfoPackage to specify the selection parameters for loading data into the PSA. In the transformation, you determine how the fields of the are assigned to the BI InfoObjects. Data transfer processes facilitate the further distribution of data from the PSA to other targets. The rules that you set in the transformation are applied here.

Overview of Object Types

A DataSource cannot exist simultaneously in both object types in the same system. The following table provides an overview of the (transport-relevant) metadata object types. The table also includes the object types for DataSources in SAP source systems:

DataSource Type


Object Type of A or M Version


Object Type of Shadow Version (Source System Independent)

SAP Source System:

Object Type of A Version

SAP Source System:

Object Type of D Version



(Shadow object delivered in its own table with release and version)



3.x DataSource


R3TR SHFS for non-replicating source systems

SHMP for replicating source systems, that is SAP source systems
(shadow object delivered in its own table with source system key)




The new DataSource concept cannot yet be used for transferring data from external systems (metadata and data transfer using staging BAPIs) or for transferring hierarchies.


We recommend that you adapt the data flow for the DataSource and the process design to the new concepts. If you want to migrate an existing data flow, first use an emulated 3.x DataSource and create a transformation for it. You can then migrate the 3.x DataSource to a DataSource and benefit from the new concepts in your scenario.

Leaving content frame

No comments: