Skip to content
DP-700 Fabric Engineer Hub

Get Started with Lakehouses in Microsoft Fabric

Get Started with Lakehouses in Microsoft Fabric

Source

Summary

This module introduces the core architectural foundation of Microsoft Fabric: the Lakehouse. A Lakehouse in Microsoft Fabric unifies the flexibility and low-cost storage of a data lake with the transactional consistency, schema enforcement, and relational querying capabilities of a data warehouse.

Key Architectural Concepts

1. OneLake Storage Architecture

  • One Data Lake for the entire tenant: OneLake is built on Azure Data Lake Storage (ADLS) Gen2 and automatically provisions hierarchical storage for all Fabric workspaces.
  • Tables vs. Files Section:
    • Tables/: Managed area where data is stored in open Delta Lake (Parquet) format. Any table in this folder is automatically registered in the Fabric Metastore and exposed via the Lakehouse SQL analytics endpoint.
    • Files/: Unmanaged raw storage area for files of any format (CSV, JSON, Parquet, images, audio, uncompressed logs). Files can be transformed and written into Tables/ via Spark.

2. Lakehouse Multi-Engine Access

Every Fabric Lakehouse automatically provides two complementary endpoints without copying data:

  1. Lakehouse Spark Endpoint: Full read/write access via Apache Spark / PySpark notebooks for complex transformations, machine learning, and data engineering.
  2. SQL Analytics Endpoint: High-performance, read-only T-SQL query engine that reads Delta tables directly. Supports standard T-SQL queries, views, table-valued functions, and Row/Column-level security.

3. V-Order Optimization & Direct Lake Mode

  • Fabric Delta parquet files are written with V-Order sorting enabled by default, optimizing in-memory compression and lightning-fast scan times for the Power BI Direct Lake engine and SQL engine.

Hands-On Practices & Lab Exercises

  • Creating a Fabric workspace with Lakehouse items.
  • Uploading raw CSV and JSON files to the Files/ container.
  • Loading files into managed Delta tables using the Load to Tables UI feature.
  • Querying Delta tables using the SQL analytics endpoint.

Exam Traps & Gotchas

[!WARNING]

  • SQL Analytics Endpoint is strictly Read-Only: You cannot execute INSERT, UPDATE, DELETE, or DROP TABLE from the SQL analytics endpoint. Write operations must occur via Spark notebooks, Dataflows Gen2, or Data Factory pipelines.
  • Files vs Tables Discovery: Files placed in the Files/ directory are NOT visible to the SQL analytics endpoint or Power BI Direct Lake mode until loaded into Delta tables under Tables/.