An end-to-end analytics SaaS platform that unifies data engineering, data integration, data warehousing, data science, real-time analytics, and business intelligence in a single environment.
OneLake
Built-in unified data lake
Fabric’s central storage layer. OneLake stores all organizational data in open formats and provides a single source of truth across Fabric workloads.
Lakehouse
Workspace → New → Lakehouse
Combines data lake flexibility with warehouse-style structure. Stores files and Delta tables for analytics, notebooks, SQL queries, and Power BI reporting.
Warehouse
Workspace → New → Warehouse
A fully managed SQL-based analytics warehouse optimized for structured relational querying and reporting workloads.
Data Factory
Data Factory experience
Fabric’s data integration layer for creating pipelines, orchestrating data movement, and scheduling transformations across multiple systems.
Real-Time Analytics
Eventstream / KQL Database
Handles streaming and high-frequency telemetry data using Kusto Query Language (KQL) for fast querying and monitoring.
Power BI
Integrated within Fabric
Provides visualization, semantic modeling, dashboards, and enterprise reporting capabilities directly within Fabric workspaces.
Workspaces & Storage
Workspace
Fabric portal → Workspaces
Logical container for organizing Fabric assets such as Lakehouses, Warehouses, Pipelines, Reports, and Notebooks.
Capacity
F SKU (F2, F4, F64, etc.)
Dedicated compute resources assigned to Fabric workloads. Determines performance, concurrency, and feature availability.
Shortcuts
Lakehouse → New Shortcut
Creates virtual references to external storage locations without copying the data into OneLake.
Delta Tables
Tables stored in Delta format
Transactional storage format supporting ACID compliance, schema evolution, and scalable analytics.
Files Section
Lakehouse → Files
Stores raw and semi-structured files such as CSV, JSON, Parquet, and text files before transformation.
Tables Section
Lakehouse → Tables
Contains structured Delta tables optimized for SQL analytics and Power BI consumption.
Data Engineering
Notebook
Workspace → New → Notebook
Interactive Spark-based environment supporting Python, SQL, Scala, and Spark operations for large-scale data engineering.
Apache Spark
Spark runtime in Fabric
Distributed processing engine used for large-scale transformations, machine learning, and notebook execution.
PySpark
Python API for Spark
Allows scalable distributed data processing using Python syntax within Fabric notebooks.
DataFrame
spark.read.format(...)
Distributed tabular data structure used in Spark for transformations, filtering, aggregation, and joins.
Save as Table
df.write.saveAsTable()
Writes Spark DataFrame results directly into a managed Lakehouse Delta table.
Semantic Link
Semantic model integration
Enables notebooks and machine learning workloads to directly interact with Power BI semantic models.
Data Factory & Pipelines
Pipeline
Data Factory → New Pipeline
Workflow orchestration engine used to automate ingestion, transformation, and movement of data.
Copy Activity
Pipeline → Copy Data
Moves data from source systems into Fabric destinations such as Lakehouses and Warehouses.
Dataflow Gen2
New → Dataflow Gen2
Modern Power Query-based ETL tool inside Fabric that supports scalable cloud-native transformations.
Schedule Pipeline
Pipeline → Schedule
Configures recurring execution of data workflows at specified intervals.
Parameters
Pipeline → Parameters
Dynamic values passed into pipeline activities for reusable and configurable workflows.
Monitor Hub
Fabric → Monitor
Centralized interface for viewing execution history, failures, durations, and refresh status across Fabric workloads.
SQL & Warehousing
SQL Endpoint
Lakehouse → SQL analytics endpoint
Automatically generated SQL interface that enables T-SQL querying against Lakehouse tables.
T-SQL
SELECT * FROM table_name
Primary SQL dialect used in Fabric Warehouses and SQL Endpoints for querying and transformations.
Stored Procedure
CREATE PROCEDURE
Reusable SQL routines that encapsulate logic and automate database operations.
View
CREATE VIEW
Virtual table defined by a SQL query to simplify reporting and abstract complexity.
CTAS
CREATE TABLE AS SELECT
Creates a new table from the output of a SQL query. Common for staging and transformations.
Star Schema
Fact + Dimension tables
Recommended modeling pattern for analytics workloads. Optimizes performance and simplifies reporting.
Real-Time Analytics
Eventstream
Real-Time Analytics → Eventstream
Captures and routes streaming data from sources such as IoT devices, applications, and logs.
KQL Database
New → KQL Database
High-performance analytics engine optimized for log analytics, telemetry, and time-series workloads.
Kusto Query Language (KQL)
StormEvents | take 10
Read-optimized query language designed for fast filtering, aggregation, and analytics on streaming data.
Real-Time Dashboard
KQL Dashboard
Interactive dashboard for monitoring live streaming data and operational metrics.
Streaming Data
Continuous ingestion
Data arriving continuously in near real time from devices, APIs, sensors, or applications.
Power BI Integration
Semantic Model
Dataset layer in Fabric
Business-friendly data model used for Power BI reporting, DAX calculations, and reusable analytics.
Direct Lake
Direct access to OneLake
High-performance mode allowing Power BI to query Delta tables directly from OneLake without import refreshes.
Import Mode
Cached dataset
Loads data into memory for fast query performance. Requires scheduled refresh to stay current.
DAX
Measures and calculated columns
Expression language used for calculations and analytics in Power BI semantic models.
Report
Power BI Report
Interactive analytical canvas containing charts, filters, KPIs, and drill-through capabilities.
Dashboard
Pinned report visuals
High-level monitoring interface created from pinned report visuals and KPIs.
Security & Governance
Microsoft Purview
Governance integration
Provides data cataloging, lineage, sensitivity labels, and governance across Fabric assets.
Row-Level Security (RLS)
Manage Roles
Restricts access to specific rows of data based on the logged-in user.
Sensitivity Labels
Confidential / Highly Confidential
Classifies and protects sensitive organizational data across Fabric and Microsoft 365.
Data Lineage
Workspace lineage view
Tracks upstream and downstream dependencies between datasets, reports, pipelines, and tables.
Workspace Roles
Admin / Member / Contributor / Viewer
Controls permissions and collaboration capabilities within Fabric workspaces.
Best Practices
Use Medallion Architecture
Bronze → Silver → Gold
Layer raw, cleaned, and business-ready data into separate stages for maintainability and governance.
Optimize Delta Tables
VACUUM / OPTIMIZE
Maintain Delta table performance by compacting files and cleaning obsolete data.
Separate Storage & Compute
OneLake + Capacity
Use OneLake for centralized storage while scaling compute independently using Fabric capacities.
Prefer Direct Lake
Power BI Direct Lake mode
Reduces refresh overhead and improves report performance by querying OneLake directly.
Implement Naming Standards
Consistent table/report naming
Use structured naming conventions for workspaces, tables, pipelines, and reports to improve governance.