Concept card

Data Lineage

As of 2026-09-27

Edited by Cédric Mary, sole editor and publication director — 27 years in the SAP ecosystem. How these pages are produced, and where automation is used.

What is Data Lineage?

Datasphere auto-captures lineage for everything built inside it, but external sources still need manual Catalog-API entries — skip that and the lineage graph quietly stops being true the day an auditor asks for it.

What it is

Data lineage = the documented flow of every dataset from origin to consumption. Source system → ingestion (Replication Flow / Data Flow) → modeling layer (views) → semantic layer (analytic models) → consumption (SAC stories, Joule answers, Excel pivots). Without lineage, every audit is archaeology; every change is a coin flip; every regulatory inquiry takes weeks.

Why lineage matters in 2026.

GDPR Art. 30 — record of processing activities requires data flow documentation.
EU AI Act Art. 10 — training data governance for high-risk AI systems requires lineage.
CSRD reporting — sustainability disclosures require auditable derivation paths.
Practical operations — when a metric changes unexpectedly, lineage tells you which view, which transformation, which source caused it.

Datasphere's auto-captured lineage. Every view, analytic model, and Replication Flow auto-registers its sources + targets in the Catalog. The lineage graph navigates: click any object, see upstream sources and downstream consumers. Junior consultants don't even know it's there; senior consultants treat it as a deliverable.

Senior craft signal — manual lineage supplement. Auto-capture covers the DSP-internal flow. External sources (legacy systems, third-party feeds, manual CSVs) need manual lineage entries via the Catalog API. A senior consultant builds a lineage-completion gate into go-live: every analytic model traces back to a named source system, with a documented owner and refresh SLA.

Lineage at story time. SAC stories that bind to live DSP analytic models inherit lineage automatically. Imported flat models break the lineage chain — yet another reason imported flat is the last-resort pattern.

Why it matters

GDPR Art. 30, EU AI Act Art. 10, and CSRD reporting all now require an auditable data-flow trail, not just a working dashboard — lineage is a compliance deliverable, not a nice-to-have.
Imported flat models break the lineage chain outright, since SAC stories only inherit lineage automatically when bound to live Datasphere Analytic Models.
Refactoring a view without updating downstream models silently breaks lineage — without quarterly hygiene reviews, the Catalog's lineage graph drifts from reality and loses the team's trust.

Key points

Datasphere auto-captures DSP-internal lineage
Manual entries needed for external sources
Mandatory pre-cert gate at go-live
Refactor lineage in same commit as code refactor
Quarterly hygiene catches drift
Art. 10 evidence requires lineage documentation
Imported flat models break the chain
Owner + SLA + source URL per external feed

Terms used on this page

Data lineage: Documented data flow from source to consumption
Auto-captured lineage: Datasphere built-in lineage for internal objects
Manual lineage entry: Catalog API call to register external sources
Lineage gate: Pre-certification check that all sources are documented
Lineage drift: Stale lineage after refactor without update
Hygiene review: Quarterly cadence to detect and fix drift
Art. 10: EU AI Act training data governance article
Lineage navigation: UI to traverse upstream/downstream from any object

Sources

Full card available to members. What the full card adds: the full decision framework · the SAP vs Snowflake / Databricks / Fabric comparison · the common pitfalls and their fix · the cheat sheet · the architecture schemas · the code blocks · the facts worth quoting.

Open in the app →

Guides that answer with this page

These guides cite this page as one of the sources their answer rests on.