🔒 Guided

Pre-launch preview. Authorised access only.

Incorrect code

Guided by A Guide to Cloud
Explore AB-900 AI-901
Guided PL-300 Domain 1
Domain 1 — Module 2 of 7 29%
2 of 26 overall

PL-300 Study Guide

Domain 1: Prepare the Data

  • Connecting to Data Sources Free
  • Storage Modes: Import, DirectQuery & DirectLake Free
  • Data Profiling & Quality Free
  • Power Query Transforms Free
  • Combining Data: Merge, Append & Queries
  • Fact Tables, Dimension Tables & Keys
  • Data Loading & Query Optimisation

Domain 2: Model the Data

  • Star Schema & Relationships
  • Date Tables & Table Properties
  • Columns vs Measures: When to Use Which
  • DAX Fundamentals Free
  • CALCULATE & Filter Context
  • Time Intelligence & Calculation Groups
  • Model Performance Optimisation

Domain 3: Visualize and Analyze the Data

  • Choosing & Formatting Visuals
  • Slicers, Filters & Interactions
  • Bookmarks, Tooltips & Navigation
  • Copilot in Power BI Reports
  • Report Pages & Paginated Reports
  • Mobile, Accessibility & Personalisation
  • AI Visuals & Pattern Discovery

Domain 4: Manage and Secure Power BI

  • Workspaces & Distribution
  • Dashboards, Alerts & Subscriptions
  • Gateways & Scheduled Refresh
  • Row-Level Security & Access Control
  • Sensitivity Labels & Governance

PL-300 Study Guide

Domain 1: Prepare the Data

  • Connecting to Data Sources Free
  • Storage Modes: Import, DirectQuery & DirectLake Free
  • Data Profiling & Quality Free
  • Power Query Transforms Free
  • Combining Data: Merge, Append & Queries
  • Fact Tables, Dimension Tables & Keys
  • Data Loading & Query Optimisation

Domain 2: Model the Data

  • Star Schema & Relationships
  • Date Tables & Table Properties
  • Columns vs Measures: When to Use Which
  • DAX Fundamentals Free
  • CALCULATE & Filter Context
  • Time Intelligence & Calculation Groups
  • Model Performance Optimisation

Domain 3: Visualize and Analyze the Data

  • Choosing & Formatting Visuals
  • Slicers, Filters & Interactions
  • Bookmarks, Tooltips & Navigation
  • Copilot in Power BI Reports
  • Report Pages & Paginated Reports
  • Mobile, Accessibility & Personalisation
  • AI Visuals & Pattern Discovery

Domain 4: Manage and Secure Power BI

  • Workspaces & Distribution
  • Dashboards, Alerts & Subscriptions
  • Gateways & Scheduled Refresh
  • Row-Level Security & Access Control
  • Sensitivity Labels & Governance
Domain 1: Prepare the Data Free ⏱ ~14 min read

Storage Modes: Import, DirectQuery & DirectLake

Understand the three storage modes in Power BI — Import, DirectQuery, and DirectLake — their trade-offs, and when to choose each one for your data model.

Import, DirectQuery, or DirectLake?

☕ Simple explanation

Think about reading a book.

Import is like photocopying the entire book and bringing it home. You can read it anytime, even offline — but your copy might be outdated if the author publishes new chapters.

DirectQuery is like reading the book in the library. You always see the latest version, but you have to go to the library every time — and if it’s busy, you wait.

DirectLake is like the library delivering pages to your desk on demand. You get the latest content without photocopying anything, and it’s almost as fast as having your own copy — because the data lives in OneLake and Power BI pages it into memory directly from Delta tables.

Storage mode determines how Power BI stores and retrieves data for your semantic model. It’s the single most impactful architecture decision you’ll make.

Import loads data into Power BI’s in-memory VertiPaq engine. Queries are extremely fast because everything is cached locally. Trade-off: data is only as fresh as your last refresh, and large datasets consume memory.

DirectQuery sends queries to the source database in real time. Data is always current, but every visual interaction generates a query to the source — meaning performance depends on the source database’s speed and capacity.

DirectLake (requires Microsoft Fabric) reads Delta tables directly from OneLake, paging data into VertiPaq structures on demand. It combines Import-like query speed with near-real-time freshness — no scheduled refresh needed because it reads the current state of Delta tables directly.

The three modes compared

Choose your storage mode based on freshness, speed, and infrastructure requirements
FeatureImportDirectQueryDirectLake
Data freshnessSnapshot — as fresh as last refreshReal-time — always currentNear real-time — reads from OneLake directly
Query speedVery fast (in-memory VertiPaq)Depends on source databaseFast (reads Delta tables from OneLake)
Data size limitCapacity-dependent (1 GB shared, much higher on Premium/Fabric)No practical limit — data stays at sourceCapacity-dependent (OneLake storage limits apply)
DAX supportFull DAX supportMost DAX supported; some limitations vary by sourceFull DAX support
Refresh needed?Yes — scheduled or on-demandNo — always liveNo — reads current files
Source loadOnly during refreshEvery visual interaction queries the sourceMinimal — reads files, not live queries
RequiresAny Power BI licenceAny Power BI licenceMicrosoft Fabric capacity (any SKU)

When to use each mode

Kenji Tanaka at Apex Manufacturing (🏭) has different data needs across his reports:

Import — His monthly executive summary uses historical data that doesn’t change during the day. Import is perfect: fast queries, full DAX support, and the data only needs refreshing once a day.

DirectQuery — The factory floor dashboard shows live production counts. Kenji needs real-time data, so he uses DirectQuery against the factory SQL database. The trade-off: each slicer click sends a new query to SQL Server.

DirectLake — Apex recently moved their data warehouse to Microsoft Fabric’s OneLake. Kenji’s new supply chain report uses DirectLake mode — it reads Delta tables directly from OneLake, giving him near-real-time data with Import-like speed. No scheduled refresh needed.

💡 Exam tip: the DirectLake decision

DirectLake is the newest mode and appears in the April 2026 study guide for the first time. The exam will likely test:

  • When to choose DirectLake: Data already in OneLake (Fabric), need for near-real-time without DirectQuery overhead
  • When NOT to choose DirectLake: No Fabric capacity, data not in OneLake/Delta format, need for complex Power Query transformations (DirectLake has limited transform support)
  • Fallback behaviour: If DirectLake can’t serve a query directly, it falls back to DirectQuery mode against the SQL analytics endpoint

Remember: DirectLake requires Microsoft Fabric capacity (any SKU) — it’s not available with Pro or PPU licences alone. Capacity guardrails (max rows, max columns) vary by SKU.

Composite models: mixing modes

You don’t have to choose just one mode. Composite models let you mix Import and DirectQuery tables in the same model.

Riley at Coastal Fresh (🛒) imports her product catalogue (small, rarely changes) but uses DirectQuery for her sales transactions table (large, updated constantly). This gives her fast lookups on products with live data on sales.

When composite models make sense:

  • Some tables are large and need real-time data (DirectQuery)
  • Other tables are small reference data (Import)
  • You want full DAX on the imported tables while keeping live data flowing
Deep dive: Dual storage mode

In a composite model, you can set individual tables to Dual mode. Dual tables can act as either Import or DirectQuery depending on the query context. Power BI’s query engine decides which mode is fastest for each query.

Dual mode is useful for tables that are used in relationships with both Import and DirectQuery tables — typically dimension tables in a star schema.

Knowledge check

Question

What storage mode gives you the fastest queries?

Click or press Enter to reveal answer

Answer

Import mode — data is loaded into Power BI's in-memory VertiPaq engine, so queries don't need to contact the source database.

Click to flip back
Question

What does DirectLake require that Import and DirectQuery don't?

Click or press Enter to reveal answer

Answer

Microsoft Fabric capacity (any SKU). DirectLake reads Delta tables directly from OneLake — which is a Fabric feature. Capacity guardrails vary by SKU.

Click to flip back
Question

What is a composite model?

Click or press Enter to reveal answer

Answer

A Power BI model that mixes Import and DirectQuery tables in the same dataset. Commonly used to import small reference tables while using DirectQuery for large fact tables.

Click to flip back
Question

What happens when DirectLake can't serve a query?

Click or press Enter to reveal answer

Answer

It falls back to DirectQuery mode, sending the query to the SQL analytics endpoint in Fabric. This is automatic but slower than native DirectLake.

Click to flip back
Knowledge Check

Kenji's factory floor dashboard needs to show real-time production counts from a SQL Server database. Which storage mode should he use?
Knowledge Check

Which storage mode reads Delta tables directly from OneLake with on-demand paging into VertiPaq?

🎬 Video coming soon

Next up: Data Profiling and Quality — how to spot bad data before it breaks your model.

← Previous

Connecting to Data Sources

Next →

Data Profiling & Quality

Guided

I learn, I simplify, I share.

A Guide to Cloud YouTube Feedback

© 2026 Sutheesh. All rights reserved.

Guided is an independent study resource and is not affiliated with, endorsed by, or officially connected to Microsoft. Microsoft, Azure, and related trademarks are property of Microsoft Corporation. Always verify information against Microsoft Learn.