SAP tableObjectROOSOURCEModuleBW_ANALYTICS

ROOSOURCE table — ROOSOURCE: DataSource metadata in the source system

ROOSOURCE stores the header metadata of a DataSource as known to the source system: technical name, logical system, application component, object version, and the delta process type assigned to it. It is the source-system counterpart of the BW-side DataSource definition and is what BW reads during replication.

This page covers what a row in ROOSOURCE actually represents, the fields worth querying, and the joins to extract-structure and generation tables. The pitfalls section focuses on the recurring mistake of treating this table as the BW definition of a DataSource when it is actually the source-system side of the story.

Published 15 Sept 2026· 900 words

What it stores

One row represents one version of one DataSource as it exists in the system that generates the extraction: technical name, the logical system it belongs to, the application component it is filed under, and which delta process (if any) has been assigned to it. Because a DataSource can exist in several object states at once, the same technical name can appear multiple times, once per OBJVERS value, as it moves from delivered content, through a modified working copy, to an activated version. BW does not maintain this table directly; it reads it when a DataSource is replicated from the source system into the BW metadata repository. If a change was made at extractor level but never replicated, ROOSOURCE in the source system already reflects it while the BW-side copy still does not.

Key fields

  • OLTPSOURCE - technical name of the DataSource
  • LOGSYS - logical system of the source system that owns this DataSource entry
  • OBJVERS - object version: A active, M modified, D delivered
  • DELTA - delta process type assigned to the DataSource at source-system level
  • APPLNM - application component the DataSource is filed under

How it joins the data model

  • ROOSOURCE-OLTPSOURCE = ROOSFIELD-OLTPSOURCE, to pull the field list and transfer structure of the DataSource
  • ROOSOURCE-OLTPSOURCE = ROOSGEN-OLTPSOURCE, to check the generation status and generated extraction program for the DataSource
  • ROOSOURCE-OLTPSOURCE = RSMONICDP-DATASOURCE (join on the DataSource name), to correlate the metadata definition with actual extraction monitor entries for a given request

How to read it safely

Always restrict on OBJVERS first, normally to A for the active definition, otherwise a query returns the delivered, modified and active copies stacked together and looks like duplication. Restrict on LOGSYS next whenever more than one source system feeds the same BW system, since two unrelated source systems can register a DataSource of the same technical name. The table itself is small, generally a few thousand rows per system, so it is never a performance risk; the risk is entirely interpretive, not volumetric. Do not assume the row seen in the BW system's own copy of this table matches what actually sits in the real source system until a replication has been triggered.

How to prove it in the data

Symptom: an InfoPackage or DTP that should pull delta only keeps behaving like a full load. Select from ROOSOURCE where OLTPSOURCE equals the DataSource name, OBJVERS equals A, and LOGSYS equals the source system in question. Read the DELTA field. If it is blank or set to a generic full-extraction value, the DataSource genuinely has no delta process defined at source, and the fix belongs in the extractor or generic extraction configuration, not in the BW load process.

ECC vs S/4HANA

Where classic DataSources are still used, including generic DataSources built on views or function modules, the source system continues to maintain this table in essentially the same shape, and BW or BW/4HANA continues to read it on replication. Where extraction has moved to ODP-based access on CDS views or extractors exposed through the operational data provisioning framework, the relevance of this particular table diminishes because the delta and field metadata are increasingly driven by the CDS annotation layer rather than by the classic DataSource shell, though a compatible entry can still appear once the object is exposed as a DataSource.

Common pitfalls

  • Reading ROOSOURCE in the BW system and treating it as current when the DataSource was changed at the source and never replicated; the BW-side entry is a snapshot as of the last replication, not a live mirror.
  • Ignoring OBJVERS and pulling all rows for a DataSource, then reporting the modified (M) version's delta setting as if it were active; only the A version is live until it has been activated.
  • Assuming the DELTA field guarantees that delta extraction functions correctly; it only records which delta process was assigned, not whether the underlying delta queue or timestamp logic is actually populating correctly.
  • Forgetting LOGSYS when a DataSource name is reused across more than one connected source system, and reading metadata belonging to the wrong system entirely.
  • Trying to fix a delta or field-selection problem by editing this table directly; it is populated by DataSource generation and replication processes, and manual changes are overwritten or ignored on the next replication.

Whose problem this is

Day-to-day ownership sits with the BW extraction or data integration consultant configuring the load. When the question is about what delta process is technically possible for a given business object, that question belongs to whoever owns the source application (FI, MM, SD, or a custom extractor developer), since the delta capability is defined there, not in BW.

Related SAP objects

Reviewed pages this object connects to in the ERPClimb knowledge graph.

Source: ERPClimb — https://erpclimb.com/sap-tables/roosourceERPClimb is an independent platform and is not affiliated with SAP SE. Reference pages are written and reviewed by SAP consultants for learning and troubleshooting.