What is a Space?

A Space is an isolated, self-contained workspace within a Datasphere tenant. Think of it as a dedicated project room — it has its own storage allocation, compute resources, users, connections, and schema. Different business units, projects, or environments (Dev/QA/Prod) typically get separate Spaces.

Space Properties & Configuration

Space Types

Cross-Space Sharing

Objects in one space can be shared to other spaces. The sharing is always read-only — the consuming space cannot modify the shared object. This enables a hub-and-spoke model where a central data space houses master data that other project spaces consume.

Space SQL Schema

Every Space is backed by a HANA Cloud schema with the same name as the Space ID. You can expose this via Open SQL Schema to allow external JDBC/ODBC connections from tools like DBeaver, SAP HANA Studio, or Tableau.

Tenant-Level Roles

Space-Level Roles

Custom Roles

Beyond predefined roles, Datasphere supports custom roles — combinations of privileges that can be tailored for specific job functions. For example, a role that allows creating views but not connections, or reading certain spaces but not others.

User Types

Local Tables

A Local Table is a table physically stored inside Datasphere's HANA Cloud schema. Data is written into this table and persisted in Datasphere. It is the only table type you can write data into.

• Created by: Defining structure manually or importing CSV/data flow output
• Storage: Consumes your Space's disk quota
• Editable: Yes — you can add/delete rows, update data
• Used for: Planning data write-back, intermediate staging, master data uploads, target of Data Flows

Local Table — Record Flags

Local tables support technical delta tracking. Each row can have a record flag indicating its change type:

Remote Tables

A Remote Table is a virtual representation of a table from a connected source system. The data physically stays at the source — Datasphere creates a metadata pointer to it. When you query a remote table, the query is pushed down to the source system.

• Created by: Importing table metadata from a connected source
• Storage: Zero — data stays at source
• Editable: No — read-only
• Performance: Depends on source system and network latency
• Used for: Federation (accessing live source data), initial exploration before deciding to replicate

Replica Tables

A Replica Table is a remote table that has been replicated — its data has been copied into Datasphere's local HANA storage. It is still "owned" by the remote source but has a local copy for performance.

Table Semantic Types

Every table (and view) in Datasphere has a Semantic Usage setting that defines its role in the data model:

Partitioning

Local tables support table partitioning to optimize query performance for large datasets. Partitioning splits the table into smaller physical segments based on a partition key.

• Hash partitioning — distributes rows evenly across partitions
• Range partitioning — by date/value range (e.g. one partition per year)
• Queries that filter on the partition key only scan relevant partitions — significantly faster for large tables

View Types

Graphical View Nodes

SQL View — Key Syntax Features

Datasphere SQL Views support SAP HANA SQL dialect (Standard SQL mode). Key supported features:

-- Basic join
SELECT "Products"."PRODUCTID", "ProductTexts"."SHORT_DESCR"
FROM "Products"
INNER JOIN "ProductTexts" 
  ON "Products"."PRODUCTID" = "ProductTexts"."PRODUCTID"
  AND "ProductTexts"."LANGUAGE" = 'EN';

-- Window function
SELECT SALESORDERID, AMOUNT,
  SUM(AMOUNT) OVER (PARTITION BY CUSTOMERID ORDER BY ORDERDATE) AS RUNNING_TOTAL
FROM "SalesOrders";

-- CASE expression
SELECT PRODUCTID,
  CASE WHEN PRICE > 1000 THEN 'Premium'
       WHEN PRICE > 100  THEN 'Mid-range'
       ELSE 'Budget' END AS PRICE_SEGMENT
FROM "Products";

Input Parameters

Views can define Input Parameters — runtime variables that filter data when the view is queried. For example, a LANGUAGE parameter that limits text records to one language.

• Defined in the view properties panel
• SAC prompts the user for input parameter values when opening a story based on such a view
• Can have default values
• Useful for multi-language models, fiscal year variants, company code filters

View Persistency

By default views are virtual — executed at query time. For complex views with many joins or aggregations, you can enable Persistency to materialize the view result as a snapshot in HANA storage.

Expose for Consumption

The Expose for Consumption toggle (found in Model Properties) makes a view or analytical dataset available for SAC live connections and external JDBC/ODBC access. When OFF, the object is internal-only and invisible to SAC.

INNER JOIN

Returns only rows that have a match in both tables. Non-matching rows from either side are dropped.

SELECT P."PRODUCTID", P."PRICE", T."SHORT_DESCR"
FROM "Products" P
INNER JOIN "ProductTexts" T ON P."PRODUCTID" = T."PRODUCTID";

LEFT JOIN (LEFT OUTER JOIN)

Returns all rows from the left table plus matching rows from the right. Non-matching right rows → NULL.

SELECT P."PRODUCTID", P."PRICE", T."SHORT_DESCR"
FROM "Products" P
LEFT JOIN "ProductTexts" T ON P."PRODUCTID" = T."PRODUCTID";

RIGHT JOIN

Returns all rows from the right table plus matching rows from the left. Non-matching left rows → NULL. Rarely used — usually rewritten as LEFT JOIN with tables swapped.

FULL OUTER JOIN

Returns all rows from both tables. NULLs fill in where there is no match on either side.

CROSS JOIN

Cartesian product — every row from left × every row from right. No ON condition. Rows = left_count × right_count.

-- 100 products × 12 months = 1,200 rows
SELECT P."PRODUCTID", M."MONTH"
FROM "Products" P CROSS JOIN "MonthDim" M;

Join Cardinality

In Datasphere's graphical view join node, you must specify cardinality — the expected relationship between the two tables. This helps the HANA optimizer generate better query plans.

Multi-column Join Conditions

Joins can have multiple conditions using AND. A common example in Datasphere's sample data:

-- Ensures product text language matches category text language
INNER JOIN "ProductTexts" AS T 
  ON T."PRODUCTID" = P."PRODUCTID" 
  AND T."LANGUAGE" = C."LANGUAGE"

Without the LANGUAGE condition, each product row would join to all language variants of its text — multiplying rows unintentionally.

Join Ordering Impact

In nested joins (chained in graphical view), the order matters for two reasons:

• A table can only reference columns from tables already joined above/before it in the chain
• Join order affects intermediate result set sizes, which impacts performance
• Moving a table to an earlier join point may lose reference to columns it needs (like LANGUAGE from a table joined later)

What is an Analytic Model?

An Analytic Model sits on top of an Analytical Dataset (your fact view) and assembles the complete analytical structure — dimensions, attributes, hierarchies, measures, filters, and currency settings. It is the primary object that SAC consumes for reports and stories.

Building Blocks

Measures — Types

Analytic Model vs Analytical Dataset — when to use which

Flow Type Comparison

Replication Flow

A Replication Flow connects to an external source and replicates tables — either as a one-time full load or continuously with delta (changes). It is the recommended way to bring source system data into Datasphere at scale.

• Initial Load: Full copy of the source table
• Delta: Continuous or scheduled replication of changes (INSERT/UPDATE/DELETE) using the source's change capture mechanism
• Supported sources: S/4HANA, BW, ABAP systems, SAP HANA, ODP sources (BW extractors)
• Multiple tables can be added to one Replication Flow — batch replication management
• Target is always a Local Table (auto-created with matching structure)

Data Flow

A Data Flow is a visual ETL pipeline — similar to a mapping in SAP Data Services or an SSIS package. It has sources, transformation operators, and a target local table.

Data Flow Operators

Transformation Flow

A Transformation Flow is a lightweight, SQL-based flow that reads from one or more local tables or views and writes into a local table. It supports UPSERT (update if exists, insert if not) and Full Load modes.

Scheduling

All flow types support scheduling via the built-in Datasphere scheduler:

• One-time execution
• Recurring: Hourly, Daily, Weekly, Monthly
• Cron-based for precise scheduling
• Flows can be chained — a completed replication triggers a transformation flow

Task Chains

A Task Chain is an orchestration object that sequences multiple flows, view persistency refreshes, or other tasks in a defined order with dependency handling — similar to a BW Process Chain.

Connection Architecture

Connections are defined at the Space level — each space maintains its own set of connections. A connection can be shared from one space to another (like other objects). Connections are created by the Space Administrator or DW Integrator role.

SAP Source Connections

Cloud & Third-Party Connections

Connection Status Monitoring

Each connection shows a connection status — you can test it at any time from the connection editor. Common status values:

• OK — connection is healthy and reachable
• Error — credentials invalid, network unreachable, or DP Agent offline
• Warning — connected but with degraded performance or partial access

Secure Credential Storage

Connection credentials (usernames, passwords, API keys) are stored encrypted in Datasphere's secure store — they are never displayed in plain text after initial entry. Service accounts (technical users) should be used rather than personal credentials.

Federation (Virtual Access)

In federation mode, data stays at the source. Datasphere creates a virtual (remote) table that, when queried, sends the query to the source system and returns results in real time. There is zero data copy and zero storage consumed in Datasphere.

• Latency: Network + source system query time (can be slow)
• Freshness: Always live — reflects the source at query time
• Good for: Large infrequently queried tables, operational reporting where freshness > speed, initial exploration
• Risk: Adds load to source system; source outage = no data in Datasphere

Replication (Physical Copy)

In replication mode, data is physically copied from the source into Datasphere's HANA Cloud storage. The replica table contains an up-to-date local copy of source data, refreshed on schedule or via delta capture.

• Latency: Fast — query hits local HANA
• Freshness: As of last replication (may be seconds to hours depending on mode)
• Good for: Reporting, SAC stories, joined analytics, high-frequency queries
• Cost: Consumes storage quota in the Space

Replication Modes

Switching Between Modes

For a remote table, you can switch between Federation and Replication at any time in the remote table editor — under Data Access → Remote / Replicated. Switching to Replicated triggers an initial load to copy data into HANA.

Federation Decision Framework

Why DP Agent?

Datasphere runs in the cloud. On-premise systems (S/4HANA on-premise, BW, Oracle, SQL Server) are behind firewalls and not directly reachable from the internet. The DP Agent is a lightweight Java agent installed on a server in your on-premise network that:

• Initiates an outbound HTTPS connection to Datasphere (no inbound firewall rules needed)
• Acts as a proxy — Datasphere sends queries/commands to the agent, which executes them locally against the source
• Returns results securely over the same HTTPS tunnel

DP Agent Architecture

DP Agent Setup Steps

DP Agent vs HANA Cloud Connector

Why a Semantic Layer?

Data Builder views are technically correct but often use technical column names like KBETR, WAERS, MENGE_ST. A business analyst working in SAC needs to see Rate, Currency, Quantity. The Business Layer bridges this gap — it adds business names, descriptions, relationships, and KPI logic on top of the technical model.

Business Layer Objects Overview

Associations

In the Business Layer, Associations define relationships between Business Entities — similar to foreign key relationships but declared at the semantic level rather than enforced at the database level.

For example: a SalesOrders business entity → associated to BusinessPartners via PARTNERID. When SAC queries through this model, HANA automatically resolves the join at query time.

Creating a Business Entity

You create a Business Entity on top of an existing Data Builder view or table. The entity inherits all columns from the source and you annotate them:

• Business Name: User-friendly label (e.g. "Sales Amount" instead of "NETWR")
• Description: Full definition visible in catalog and SAC
• Key: Mark columns that uniquely identify a record
• Attribute Type: Text, ID, Amount, Quantity, Date, etc.
• Semantic Type: Fact, Dimension, Measure

KPIs (Key Performance Indicators)

Business Entities support defining KPIs — business metrics with target values, thresholds, and trend direction. KPIs appear in the Data Catalog and can be surfaced in SAC dashboards.

Calculated Attributes

Beyond simple column mappings, Business Entities support Calculated Attributes — new columns derived by formula at the semantic level. Similar to a calculated column in a view, but defined in the business layer without modifying the underlying technical model.

Consumption Model Structure

A Consumption Model takes one Fact Business Entity as its center and associates multiple Dimension Business Entities to it — forming a semantic star schema. Additional properties like filters, variables, and currency settings can be defined at this level.

Consumption Model vs Analytic Model

Perspectives

A Perspective is a subset of a Consumption Model tailored for a specific audience or use case. For example, a single "Sales" Consumption Model can have:

• A Regional Manager Perspective — pre-filtered to the manager's region, showing only regional KPIs
• A Finance Perspective — showing revenue and cost measures, hiding operational metrics
• A Product Analysis Perspective — focused on product dimension drill-down

Perspectives reduce the need to build multiple separate models for different audiences.

What gets cataloged automatically?

Every object created in Datasphere is automatically indexed by the catalog — no manual registration needed:

Catalog Asset Page

Every cataloged asset has a detail page showing:

• Overview: Technical name, business name, type, space, owner, dates
• Schema/Structure: Columns with types, keys, measures
• Lineage: Visual diagram of upstream sources and downstream consumers
• Impact: List of objects that depend on this asset
• Tags: Classification labels attached by data stewards
• Glossary Terms: Business definitions linked to columns
• Data Quality: Quality score and rule results
• Sample Data: Preview of actual data (permission-dependent)

Business Glossary

The Glossary is a curated dictionary of business terms. Each term has a name, definition, owner, status (Draft/Approved/Deprecated), and can be linked to specific columns in catalog assets.

Tags & Classifications

Tags are free-form labels for organizing and filtering catalog assets. Classifications are structured tags (hierarchical) for formal governance programs like GDPR or data sensitivity classification.

Data Quality Rules

Data quality rules run on local tables and produce a Data Quality Score (0–100%). Rules are defined on specific columns:

• Completeness: % of non-null values in a column
• Uniqueness: % of unique values (identifies duplicates)
• Validity: % matching a defined pattern or range (e.g. date format, positive amounts)

Lineage Analysis (Upstream)

Lineage traces data backward through the dependency chain — from the consumer all the way to the original source system. The lineage diagram shows every table, view, flow, and model that data passed through.

This answers the question: "Where exactly does the Revenue figure in my SAC report come from?"

Impact Analysis (Downstream)

Impact analysis looks forward — given an object, what downstream objects depend on it? This answers: "If I change or delete this table, what reports/views/models will be affected?"

Column-Level Lineage

Datasphere provides column-level lineage — not just object lineage. You can trace a specific column (e.g. "Gross Margin %") through every transformation and calculation to its source column(s). This is critical for audit and data certification.

Cross-Space Lineage

Lineage tracks objects across Spaces. If Space A's table is shared to Space B and used in a view there, the lineage diagram shows the cross-space dependency. This provides full tenant-wide visibility.

Lineage from External Sources

For objects originating from external connections (S/4HANA, BW, cloud storage), lineage shows the source system as the origin node — even though it's outside Datasphere. The connection metadata enables this cross-system tracing.

Security Layers

Data Access Controls (DAC) — Row-Level Security

A Data Access Control is a criteria-based filter automatically applied to queries based on the logged-in user's assignments. It restricts which rows of data a user can see, transparently — the user doesn't know the filter exists.

DAC Types

DAC Assignment

DACs are attached to Analytical Datasets or Analytic Models. When SAC queries the model, Datasphere automatically appends the user's access filter before executing the query — no SAC-side configuration needed.

IDP Integration (SSO)

Datasphere integrates with any SAML 2.0-compatible Identity Provider (IdP) — Azure AD, Okta, SAP IAS (Identity Authentication Service). Users authenticate through their corporate SSO; Datasphere maps the IdP user to a Datasphere user account.

Audit Logging

Datasphere maintains audit logs for security-relevant actions — user logins, data access, configuration changes, connection usage. Audit logs can be forwarded to SAP Cloud ALM or SIEM tools for compliance monitoring.

Setting Up the Live Connection

What SAC Can Consume

Query Flow

When an SAC story user opens a chart:

1. SAC translates the chart definition into an OLAP query
2. Query is sent to Datasphere via the live connection (HTTPS)
3. Datasphere's HANA engine executes the query — applying DAC filters, join logic, measure calculations
4. Result set is returned to SAC and rendered in the chart
5. No data is cached or stored in SAC — every interaction re-queries Datasphere

Architecture Overview

In the traditional BPC architecture, planning data was stored inside BPC's own proprietary storage. The modern Datasphere+SAC architecture separates storage from UI:

• Storage layer: Datasphere Local Tables (HANA Cloud)
• Modeling layer: Datasphere views + Analytic Models
• Planning UI: SAC Planning models and Input Tasks
• Write-back: SAC writes plan data directly into Datasphere Local Tables

Setting Up Planning on Datasphere

Datasphere + SAC vs BPC Architecture

Data Actions in this Architecture

SAC Data Actions still work in Datasphere-backed planning models. They execute against the Datasphere Local Table and can perform calculations, copy versions, allocate values across periods — same as BPC Script Logic but defined in SAC.

What is BW Bridge?

BW Bridge is a special component of Datasphere that hosts a BW-compatible runtime. It allows organizations to lift and shift existing BW/4HANA objects into Datasphere without full re-development, while gradually migrating them to native Datasphere objects over time.

What BW Bridge Supports

BW Bridge Migration Approach

How It Works

When a measure is tagged with a currency column (e.g. transaction currency WAERS), Datasphere can automatically convert amounts to a target currency at query time using exchange rates stored in the system.

Exchange Rate Tables

Datasphere uses the standard SAP exchange rate tables, typically replicated from S/4HANA:

Currency Conversion Configuration

On a measure in an Analytic Model, you configure:

• Source currency column: The column containing the transaction currency code (e.g. WAERS)
• Target currency: Fixed (e.g. always EUR) or variable (SAC user chooses at runtime)
• Exchange rate type: M (monthly average), B (bank buying), S (bank selling), or custom
• Date for rate lookup: Fixed date, or dynamic (e.g. posting period date column)

Performance Levers

Monitoring Tools

• Space Monitor: CPU, memory, disk usage per Space. Identify resource-intensive flows or queries.
• Data Integration Monitor: Status and logs for all Replication Flows, Data Flows, and Task Chains. Shows run duration, row counts, errors.
• Database Explorer (HANA): Open via Datasphere to run SQL queries, inspect execution plans, view column statistics.
• SAC Performance Monitor: In SAC, story-level query performance metrics showing which queries are slow.

Common Performance Anti-Patterns

Open SQL Schema

Every Space can optionally expose an Open SQL Schema — a direct HANA SQL access point to the space's objects. External tools connect via standard JDBC/ODBC.

• Query views and tables directly using HANA SQL
• Compatible with SAP HANA Studio, DBeaver, Tableau, Power BI, Excel (via ODBC)
• Read-only — cannot INSERT/UPDATE/DELETE via Open SQL
• Access is subject to Datasphere role/DAC restrictions

-- Connect via JDBC, then query a Datasphere view
SELECT "PRODUCTID", "SHORT_DESCR", "PRICE"
FROM "MY_SPACE"."ProductCatalogView"
WHERE "PRODCATEGORYID" = 'LT';

OData API

Datasphere exposes OData v4 endpoints for consuming data programmatically. When an Analytical Dataset or Consumption Model has "Expose for Consumption" enabled, it can be accessed via OData.

GET https://<tenant>/api/v1/datasphere/spaces/MY_SPACE
    /objects/SalesAnalyticModel
    ?$select=PERIOD,AMOUNT&$filter=YEAR eq '2024'&$format=json

Datasphere REST API

Datasphere provides a Management REST API for administrative operations — automating space creation, user management, flow execution, and catalog operations. Useful for CI/CD pipelines and automated deployment.

HDI Containers

For advanced SAP HANA developers, Datasphere supports connecting HDI (HANA Deployment Infrastructure) Containers — allowing deployment of custom HANA database objects (calculation views, procedures, synonyms) that can be consumed in Datasphere alongside native objects.

Most Asked Comparison Questions

Local Table vs Remote Table vs Replica

Replication Flow vs Data Flow vs Transformation Flow

Analytic Model vs Consumption Model

Federation vs Replication

Key Terms to Know Cold

Common Interview Questions

1. What is the difference between a remote table and a replica table in Datasphere?
2. When would you use a Data Flow vs a Replication Flow?
3. How does SAC connect to Datasphere and what objects can it consume?
4. Explain the role of the Data Provisioning Agent.
5. What is the purpose of the Business Layer in Datasphere?
6. How does row-level security (DAC) work in Datasphere?
7. What is BW Bridge and what is its purpose?
8. What is the difference between Impact Analysis and Lineage Analysis?
9. How is planning write-back to Datasphere different from BPC planning?
10. What does "Expose for Consumption" do and when must it be enabled?
11. What is a Task Chain and what is its BW equivalent?
12. How would you model a multi-currency financial report in Datasphere and SAC?