📑 Table of Contents
What is SAP Datasphere
SAP Datasphere is SAP's unified data service platform providing data integration, data warehousing, data federation, and semantic modelling in a single managed service on BTP. It is the central component of SAP Business Data Cloud.
🏗 Data Integration
Replication Flows (SAP sources, CDC), Data Flows (ETL with transformations), Remote Tables (federation)
📊 Data Modelling
Graphical/SQL Views, Analytical Models, star schema design, perspectives, SQLScript functions
💼 Business Layer
Business Entities, Fact Models, Consumption Models — semantic layer for SAC live connection
🔐 Governance
Spaces isolate data and access. Data Catalog for lineage. Role-based access control.
Architecture & Layers
SAP Datasphere — layered architecture from source systems to SAC consumption
Datasphere follows a layered architecture: source systems feed data into Data Builder technical objects; Business Builder adds semantic meaning; SAC connects for consumption.
Source Systems
SAP S/4HANA, SuccessFactors, Ariba, BW, non-SAP (Salesforce, Oracle, flat files)
↓ Replication Flows / Data Flows / Remote Tables
Data Builder — IT Technical Layer
Local Tables, Remote Tables, Graphical Views, SQL Views, Analytical Models, Flows
↓ Business Builder consumes Data Builder objects
Business Builder — Business Semantic Layer
Dimension Business Entities, Fact Business Entities, Fact Models, Consumption Models, Perspectives
↓ SAC Live Connection
SAP Analytics Cloud (SAC)
Stories, Dashboards, Planning — connect to Consumption Model as live data source
Data Builder
SAP Datasphere Data Builder — object types and their relationships
Spaces
Spaces are isolated workspaces within Datasphere. Each space has its own data, users, and connections. Objects can be shared across spaces. All data engineering and modelling work happens within a Space.
Semantic Usage Property
Controls what an object can be used for in Analytical Models and Business Builder:
| Semantic Type | Behaviour |
|---|---|
| Fact / Relational Dataset | Appears in Analytical Model left panel — can be dragged onto canvas as the central fact |
| Dimension | Never dragged directly — auto-discovered via Associations after Fact is placed |
| Text | Provides language-dependent labels for dimension members |
| Hierarchy | Defines parent-child or level-based hierarchy for a dimension |
Federated vs Persisted vs Hybrid
🔴 Federated
Remote Table — live query to source at runtime. No data copied to HANA Cloud. Always current. Slower at query time.
🟢 Persisted
Local Table — data copied and stored in HANA Cloud. Fast query. Requires replication. Data has freshness latency.
🟡 Hybrid
Remote Table with replication enabled. Initially federated, then replicated to local storage for performance.
Data Builder Objects — Deep Dive
📋 Local Table
Persisted data in HANA Cloud. Target for Data Flows and Replication Flows. Physical storage object.
🌐 Remote Table
Live virtual table pointing to source system. Federated query or replication-enabled. No data copy by default.
🎨 Graphical View
Drag-drop canvas with auto-generated SQL. Visual lineage. Set semantic type. Cannot convert to SQL view.
💻 SQL View
Full HANA SQL editor. No visual lineage. Cannot convert back to graphical. Full SQL power.
📊 Analytical Model
Star-schema consumption layer. Fact + Dimensions. Used for SAC live connection (alternative to Business Builder).
📐 ER Model
Design canvas defining associations between objects. Visual relationship mapping tool.
⚙️ SQLScript Table Function
Accepts parameters, returns TABLE result. Supports multi-step logic with table variables. READS SQL DATA = read-only. Best of view (composable) + procedure (logic).
📅 Time Dimension
SAP-generated full date hierarchy Year→Quarter→Month→Day. Key = actual DATE or YYYYMM integer.
Graphical View vs SQL View
| Feature | Graphical View | SQL View |
|---|---|---|
| Design | Drag-drop canvas | SQL code editor |
| SQL generated | Auto-generated | Manual |
| Visual lineage | Yes | No |
| Conversion | Can convert to SQL view | Cannot convert back to graphical |
| Semantic type | Settable | Settable |
| Full SQL features | Limited | Full HANA SQL |
Analytical Model — Behaviour
- Left panel shows only objects with Semantic Usage = Fact or Relational Dataset
- Drag Fact onto canvas → right panel shows associated Dimensions (toggle on/off)
- Cannot add unassociated dimensions directly — must add Association on the view first
- Dimensions can be pruned (toggled off) per Analytical Model
Perspectives
Named column subsets saved on a view. Reduce canvas load, give audience-specific column visibility. Lazy — only columns in perspective are rendered at query time.
Data Flows vs Replication Flows
SAP Datasphere — Data Flow (ETL with transforms) vs Replication Flow (CDC from SAP sources)
⚙️ Data Flow — ETL Pipeline
Transform nodes: Join, Filter, Aggregate, Union, Python Script
Writes output to a Local Table
Supports SAP and non-SAP sources
Transformations Any source Complex logic
Writes output to a Local Table
Supports SAP and non-SAP sources
Transformations Any source Complex logic
📡 Replication Flow — Move Only
Move data from SAP sources with native CDC delta
Modes: Full Load / Delta / Real-time (CDC)
Always starts with a Full Load first
SAP sources Native CDC Real-time
Modes: Full Load / Delta / Real-time (CDC)
Always starts with a Full Load first
SAP sources Native CDC Real-time
Decision: Data Flow vs Replication Flow
| Scenario | Use |
|---|---|
| Transformation needed (filter, join, aggregate) | Data Flow |
| SAP source with native CDC support | Replication Flow |
| Real-time continuous streaming | Replication Flow |
| Non-SAP source | Data Flow |
| Complex use case (ingest + enrich) | RF for ingestion → staging → DF for enrichment |
Replication always starts Full Load. Delta = changed rows only (requires CDC or timestamp column). Real-time CDC = continuous streaming via SLT/ABAP delta queue. You cannot skip the initial Full Load.
Business Builder
SAP Datasphere Business Builder — semantic layer with Dimension BEs, Fact BEs, Fact Models, Consumption Models
Key difference from Data Builder: Business Builder objects are SAVED — not deployed. No Deploy button. Only Perspectives require deployment. "In Process" = incomplete configuration, not missing deployment.
Business Builder Object Hierarchy
Step 1 — Dimension Business Entities
Wrap Data Builder views. Add semantic attributes. Mark Key attribute + click Verify → Save → Ready to Use.
↓
Step 2 — Fact Business Entity
Wrap Fact view. Add measures (aggregable), attributes. Add associations to Dimension BEs → Save.
↓
Step 3 — Fact Model
Combine Fact BE + expose Dimension BEs. Predefine measures, attributes, filters for a business context → Save.
↓
Step 4 — Consumption Model
Built on Fact Model. Add dimension sources. Save → Ready to Use. SAC connects here.
↓ (Optional)
Step 5 — Perspectives
Reusable subsets of Consumption Model attributes/measures/parameters. Only BB object that requires DEPLOY.
Dimension Types
🔵 Primary Dimension
Exposed to SAC. Users filter, drill, slice by it. Added to "Expose Dimension Sources" in Fact Model. Full dimensional navigation.
🟣 Auxiliary Dimension
NOT exposed standalone. Enriches another dimension via dim-to-dim association. Its attributes surface through the parent dimension in SAC.
📝 Text Dimension
Language-dependent labels for dimension member IDs. Used for multilingual descriptions of dimension members.
Business Layer Matching — Key Terms
| Term | Definition | Key Words |
|---|---|---|
| Business Entities | Can define measures or attributes. Measures are quantifiable aggregable values. | "aggregable field" |
| Fact Models | Predefine an array of measures, attributes, and filters relevant to a specific business context. | "predefine", "business context" |
| Perspectives | Reusable configurations containing a subset of a consumption model's attributes, measures and parameters. | "subset", "reusable" |
Status Lifecycle — In Process vs Ready to Use
Critical understanding: "In Process" and "Ready to Use" are Business Builder lifecycle statuses. They indicate configuration completeness — NOT deployment state. Status cascades upward: fix Dimension BEs first, then Fact BE, then Fact Model, then Consumption Model.
⏳ In Process
Object exists but required configuration is incomplete. Cannot be used by objects above it. Not usable by SAC. Common cause: Key attribute not set or Verify not clicked on Dimension BE.
✅ Ready to Use
All required configuration complete and valid. Can be referenced by higher-level objects. Consumption Model shows its dimension sources. SAC can connect.
How to Resolve "In Process" — Fix Order
| Object | Required to reach Ready to Use |
|---|---|
| Dimension Business Entity | Set Key attribute → click Verify → Save. Key is mandatory. Without Verify, stays In Process. |
| Fact Business Entity | At least one measure added. Associations to dims defined. Dimension BEs must be Ready to Use first. |
| Fact Model | Fact BE selected (must be Ready to Use). Dims added to Expose Dimension Sources (must be Ready to Use). Save. |
| Consumption Model | Fact Model selected as source (must be Ready to Use). Save. |
| Perspective | Configured + must be Deployed (only BB object requiring deploy). |
The single most impactful fix: Open each Dimension Business Entity → Attributes tab → find the key field → mark as Key → click Verify → Save. This one step resolves most "In Process" cascades.
Consumption Models
Where Attributes Appear
Critical behaviour: Dimension attributes from the Fact Model do NOT appear in the top-level Attribute section of the Consumption Model. They are nested under each dimension node. Expand the dimension in SAC's field picker to access them.
| What | Where in Consumption Model | Where in SAC |
|---|---|---|
| Measures | Top-level Measures section | Top-level in field picker |
| Fact-level attributes | Top-level Attributes section | Top-level in field picker |
| Dimension attributes | Nested under dimension node | Expand ▶ dimension in field picker |
Why Dimension Attributes Are Nested
- Preserves star schema integrity — SAC knows which attributes belong to which dimension
- Enables drill-down, hierarchy navigation, and filter context per dimension
- Lazy loading — dimension only joined when its attributes are actually used in a chart
Associations Guide
Associations vs Joins
🔗 JOIN (main canvas)
Always executes at query time. Merges data physically. Eager — fires regardless of whether the consumer uses the joined object. Used within a single view for data combination.
🔀 Association (Associations tab)
Lazy — only fires when consumer actually uses the related object. Navigational path for Analytical/Consumption Models. Replaces explicit joins for dimensional navigation.
Association Rules — Business Builder
| Association Type | Direction | Required? | Notes |
|---|---|---|---|
| Fact → Dimension | One-way: Fact → Dim | Always — mandatory | Primary star schema join path |
| Dimension → Dimension | One-way: Child → Parent | Only when FK exists in child | Child holds the foreign key |
| Dimension → Fact | — | Never | Dimensions don't reference Facts |
| Bidirectional | — | Never needed | Associations are always one-way |
Association Context — When and Why
Use Association Context when the same two objects can be related in multiple different ways — disambiguating which join path to use for a given Fact Model.
Example: State dimension has SalesRegionID and ServiceRegionID — both pointing to the same Region dimension. Without context, Datasphere can't resolve which to use. Define two associations: one with Context="Sales" (SalesRegionID→ID) and one with Context="Service" (ServiceRegionID→ID). Each Fact Model specifies which context it activates.
Dim-to-Dim Association Rules
- Only add when child dimension physically has an FK column pointing to parent dimension
- Always from child to parent — parent never points back to child
- Auxiliary Dimensions use dim-to-dim associations — linked from primary dimension
- Not added to "Expose Dimension Sources" in Fact Model — resolved automatically via association chain
Exam Tips & Quick Reference
Critical Facts
CERT Business Builder
Objects SAVE only — no Deploy button. Exception: Perspectives require Deploy. "In Process" = incomplete config, not missing deployment.
CERT Ready to Use
Dimension BE: Key + Verify + Save. Status cascades up the chain — fix bottom first (Dim BEs → Fact BE → Fact Model → Consumption Model).
CERT Replication Flow
Always starts with Full Load. SAP sources only. Native CDC support. Delta = changed rows. Real-time = continuous via SLT.
CERT Semantic Usage
Fact/Relational Dataset = appears in Analytical Model left panel. Dimension = never dragged, auto-discovered via associations.
CERT Associations
Always one-directional. Fact → Dim (always). Dim → Dim (only when FK exists in child). Dimension → Fact: never. Bidirectional: never.
CERT Dim Attributes in SAC
NOT in top-level Attributes section. Nested under dimension node. Expand ▶ dim in SAC field picker to access them.
Data Builder Object Quick Reference
| Object | Purpose | Deploy Required |
|---|---|---|
| Local Table | Persisted storage in HANA Cloud | Yes |
| Remote Table | Federation / replication from source | Yes |
| Graphical View | Visual ETL / semantic layer | Yes |
| SQL View | Full HANA SQL, no visual lineage | Yes |
| Analytical Model | Star schema for SAC direct connection | Yes |
| Data Flow | ETL with transforms → Local Table | Yes (run) |
| Replication Flow | CDC from SAP sources → Local Table | Yes (run) |
| Dimension BE | Business semantic wrapper — attributes | No — Save only |
| Fact BE | Business semantic wrapper — measures | No — Save only |
| Fact Model | Business subject area (dim + measures) | No — Save only |
| Consumption Model | Final SAC connection point | No — Save only |
| Perspective | Subset of Consumption Model | Yes — only BB object |