Skip to content

kengio/dp-800-study-guide

BranchesTags

Repository files navigation

title DP-800 Study Guide
type project
tags
dp-800
microsoft
azure
sql
ai
certification
vector-search
rag

Microsoft Certified: SQL AI Developer Associate

DP-800: Developing AI-Enabled Database Solutions

Open-source community study guide for the Microsoft DP-800 certification

License: MIT Microsoft DP-800 Blueprint: March 2026 SQL Server 2025 Azure SQL Microsoft Fabric
T-SQL Vector Search RAG MCP Obsidian compatible PRs welcome

A community-maintained study guide for the Microsoft DP-800: Developing AI-Enabled Database Solutions certification.
Aligned to the official skills-measured list updated March 12, 2026.

Note

I used this guide to pass DP-800. ✅ These are the exact notes, cheat sheets, practice questions, and mock exams I built while preparing — refined while studying, hardened on the actual exam, and now open-sourced under MIT so you can pass too. Every section is mapped 1:1 to the official Microsoft skills-measured list. If it helped you, ⭐ the repo and pass it on.

Tip

▶ Try the practice quiz live — adaptive multiple-choice quiz, 160 questions across 3 banks (60 practice + 50 Mock 1 + 50 Mock 2, including the 5-question case-study block). Runs in your browser, progress saved locally. No install, no signup. Light/dark/auto theme + exam timer (70 min mock / 120 min real exam).

Contents

Why this guide exists

The DP-800 is Microsoft's first certification focused on building AI-enabled database solutions — vector search, embeddings, RAG, and intelligent search inside SQL Server, Azure SQL, and SQL databases in Microsoft Fabric. The official skills list is broad and changes quickly. Public study resources are scarce.

This repo is the notes that got me through it. Now it's yours.

Who this is for

  • Database developers / DBAs moving into AI-augmented workloads
  • Data engineers who want to add vector search and RAG to relational platforms
  • AI / app developers who need to talk fluently about Azure SQL, Fabric SQL, and T-SQL AI functions
  • Exam takers preparing for DP-800 specifically — every topic file maps 1:1 to the official blueprint
  • Anyone curious about how Microsoft is bringing GenAI into the database layer

You don't need to be taking the exam to get value — the guide doubles as a reference for SQL Server 2025 vector features, MCP server integration, Data API Builder, and embedding maintenance patterns.

What's covered

  • 11 topic sections mapped 1:1 to the official skills measured list
  • 7 cheat sheets for fast review (security, vector/AI, JSON, performance, T-SQL, Azure SQL config)
  • 60+ practice questions with full explanations across all three domains
  • 2 full-length mock exams (50 questions each — 45 standalone + a 5-question case study mirroring the real exam format)
  • Per-mock debrief files mapping every missed question to a topic file + cheat sheet
  • End-to-end RAG worked example — the full pipeline in ~80 lines of T-SQL
  • Final review designed to read in 20 minutes the morning of the exam
  • T-SQL code examples covering vector search, RAG, full-text, and AI integration

Exam at a glance

Detail Information
Exam ID DP-800
Full Name Developing AI-Enabled Database Solutions
Credential Microsoft Certified: SQL AI Developer Associate
Passing Score 700 / 1000
Duration 120 minutes
Cost Varies by region (commonly ~$165 USD in the US)
Renewal Annual (free Microsoft Learn assessment)
Platforms tested SQL Server (incl. 2025), Azure SQL, SQL databases in Microsoft Fabric
Language T-SQL
Blueprint date March 12, 2026

Domain weights

pie showData title DP-800 Exam Domain Distribution
    "Design and develop database solutions" : 37
    "Secure, optimize, and deploy database solutions" : 37
    "Implement AI capabilities in database solutions" : 26
Loading
Domain Weight Sections in this guide
1. Design and develop database solutions 35–40 % 01 – 04
2. Secure, optimize, and deploy database solutions 35–40 % 05 – 08
3. Implement AI capabilities in database solutions 25–30 % 09 – 11

Skills measured (high level)

Domain 1 — Design and develop database solutions (35–40 %)

Tables · indexes · columnstore · specialized tables (in-memory, temporal, external, ledger, graph) · JSON columns and indexes · constraints · sequences · partitioning · views · functions · stored procedures · triggers · CTEs · window functions · JSON functions · regex (REGEXP_LIKE, REGEXP_MATCHES, REGEXP_SPLIT_TO_TABLE, etc.) · fuzzy matching (EDIT_DISTANCE, JARO_WINKLER_DISTANCE) · graph queries with MATCH · GitHub Copilot · MCP server endpoints

Domain 2 — Secure, optimize, and deploy database solutions (35–40 %)

Always Encrypted · column encryption · Dynamic Data Masking · Row-Level Security · object-level permissions · passwordless access · auditing · Managed Identity for model endpoints · secure GraphQL/REST/MCP endpoints · isolation levels · DMVs · Query Store · Query Performance Insight · blocking and deadlocks · SQL Database Projects (SDK-style) · schema drift detection · CI/CD pipelines · Data API Builder · Azure Monitor · CDC · Change Tracking · CES · Azure Functions SQL trigger · Logic Apps

Domain 3 — Implement AI capabilities in database solutions (25–30 %)

External models · embedding maintenance (triggers, CT, CDC, CES, Azure Functions, Logic Apps, Microsoft Foundry) · chunking · embedding generation · full-text search · VECTOR data type · VECTOR_DISTANCE · VECTOR_SEARCH · VECTOR_NORMALIZE · VECTORPROPERTY · DiskANN indexes · ANN vs ENN · hybrid search · RRF (Reciprocal Rank Fusion) · RAG with sp_invoke_external_rest_endpoint

2026 updates you should know

Important

Microsoft refreshed the DP-800 skills measured on March 12, 2026. Highlights:

  • SQL Server 2025 is GA. The VECTOR data type and VECTOR_DISTANCE are generally available in SQL Server 2025 and Azure SQL Database. VECTOR_SEARCH, VECTOR_NORMALIZE, and VECTORPROPERTY are in public preview on the same platforms.
  • DiskANN vector indexes are in public preview across SQL Server 2025, Azure SQL Database, Azure SQL Managed Instance, and SQL database in Microsoft Fabric. On SQL Server 2025 also requires PREVIEW_FEATURES = ON.
  • Half-precision (16-bit) float16 vectors are in preview — halves storage for the same dimension count (the documented type cap is 1 998 dimensions).
  • MCP server endpoints (SQL Server + Fabric lakehouse) are explicitly tested.
  • Microsoft Foundry is named as a valid embedding-maintenance method alongside CDC, Change Tracking, and CES.
  • Change Event Streaming (CES) in Fabric is now in the blueprint.
  • Passwordless DB access and Managed Identity for model endpoints are explicit security requirements.
  • Schema drift detection in SQL Database Projects is now an explicit skill.

The main overview opens with the full "What's New" callout.

Getting started with Obsidian (recommended)

This guide is written in Obsidian Flavored Markdown. It renders fine on GitHub, but in Obsidian you get callouts, foldable practice-question answers, Mermaid diagrams, backlinks, and a navigable Graph View of every cross-link — which makes studying meaningfully better.

5-minute onboarding

  1. Install Obsidian (free; macOS, Windows, Linux).

  2. Clone this repo somewhere on your machine:

    git clone https://github.com/kengio/dp-800-study-guide.git
cd dp-800-study-guide

  3. Open the vault: launch Obsidian → Open folder as vault → pick the cloned dp-800-study-guide/ directory.

  4. Trust the author when Obsidian asks (the included .obsidian/ config has pre-tuned settings — line numbers, tab width, no inline titles).

  5. Open certification/dp-800-overview.md — that's your study path entry point. Press Cmd/Ctrl + O to fuzzy-find any topic.

  6. Toggle Graph View (Cmd/Ctrl + G) to see how all 11 sections cross-link — surprisingly useful for spotting weak areas.

Recommended plugins

Two are essential, the rest are quality-of-life. Install via Settings → Community plugins → Browse.

  • Obsidian Git — back up your notes and progress checkboxes to your own fork.
  • Linter — keeps your edits consistent with the project's markdown conventions.
  • Advanced Tables — auto-aligns Markdown tables as you type.
  • Codeblock Customizer (or Better CodeBlock) — line numbers, titles, copy buttons on code blocks.
  • Copilot (by logancyang) — chat with Claude / GPT-4o / Ollama inside Obsidian; Vault QA mode indexes the guide so the AI can quiz you using your actual notes.

📖 See OBSIDIAN-SETUP.md for the full setup walkthrough — plugin configuration details, Copilot Vault QA setup, recommended study prompts, and tips for using AI to generate active-recall questions from your notes.

Don't want Obsidian?

No problem. The guide also renders perfectly in:

  • GitHub — browse the files online; callouts and Mermaid diagrams render natively
  • VS Code with the Markdown All in One extension
  • Any Markdown reader that supports GFM — you'll lose callouts and Graph View, but the content is fully readable

How to use this guide

  1. Start at the main overview — it has the full study path and a progress tracker.
  2. Work through the 11 topic sections in order — each topic file is 300–600 lines with examples, comparison tables, common-mistake callouts, and exam tips.
  3. Hit the cheat sheets after each domain to consolidate.
  4. Take the practice questions — aim for 70 %+ per domain before moving on. Or drill them adaptively in the live practice quiz which surfaces what you've recently missed.
  5. Sit the two mock exams under timed conditions when you think you're close — both are available as bank options in the live quiz with a 70-minute timer.
  6. Read final-review.md the morning of the exam — it's the 20-minute scan.

Study roadmap

Pick the plan that matches the time you have. All three end with the same outcome — sitting the exam with confidence. The hours are realistic averages for an experienced T-SQL developer; double them if you're newer to relational databases or to the Microsoft stack.

🏃 4-week sprint (~30–35 hours total — ~1 hour/day)

Best for: experienced SQL developers brushing up on AI features. Tight but doable.

Week Focus Files Hours
1 Domain 1 — Design & develop 01-database-objects, 02-programmability-objects, 03-advanced-tsql, 04-ai-assisted-tools 9
2 Domain 2 — Secure, optimize, deploy 05-data-security-compliance, 06-performance-optimization, 07-cicd-database-projects, 08-azure-services-integration 9
3 Domain 3 — AI capabilities 09-models-embeddings, 10-intelligent-search, 11-rag + cheat sheets 9
4 Practice & polish Practice questions (all 3 domains) → Mock Exam 1 → review → Mock Exam 2 → final-review.md 6

🚶 8-week balanced (~55–65 hours total — ~1 hour/day, 7 days/week)

Best for: working professionals fitting study around a job. The recommended default.

Week Focus Hours
1 01-database-objects (all 5 sub-topics) + read the overview 7
2 02-programmability-objects + 03-advanced-tsql part 1 (CTEs, window, JSON) 8
3 03-advanced-tsql part 2 (regex, graph, error handling) + 04-ai-assisted-tools 7
4 Checkpoint: Domain 1 cheat sheets + Domain 1 practice questions (target 70 %+) 5
5 05-data-security-compliance + 06-performance-optimization 8
6 07-cicd-database-projects + 08-azure-services-integration 7
7 Domain 2 practice questions + 09-models-embeddings + 10-intelligent-search 8
8 11-rag + Domain 3 practice → Mock Exam 1 (timed) → review weak areas → Mock Exam 2 → final-review.md 8

🧘 12-week comprehensive (~80–100 hours total — ~1 hour/day, with weekends off)

Best for: newcomers to AI features, career changers, or anyone wanting deeper retention.

Week Focus Hours
1 01-database-objects/01–02 (tables, specialized tables) 6
2 01-database-objects/03–05 (JSON columns, constraints, partitioning) 6
3 02-programmability-objects (views, functions, procs, triggers) + first hands-on lab 7
4 03-advanced-tsql/01–03 (CTEs, JSON, regex) 7
5 03-advanced-tsql/04–05 + 04-ai-assisted-tools + Domain 1 practice questions 7
6 Checkpoint: Domain 1 cheat sheets + retake weak Domain 1 questions 5
7 05-data-security-compliance (all 5 sub-topics) 8
8 06-performance-optimization + isolation/concurrency hands-on 7
9 07-cicd-database-projects + 08-azure-services-integration 8
10 Domain 2 practice questions + 09-models-embeddings 8
11 10-intelligent-search + 11-rag (build a small RAG demo) 9
12 Domain 3 practice → Mock Exam 1 → gap-fill → Mock Exam 2 → final-review.md → exam 8

Suggested daily cadence

Weekday    (45–60 min):  Read 1 topic sub-file + work the examples in your own DB
Weekend    (90–120 min): Cheat sheet review + practice questions + flashcards
Pre-exam   (last 3 days): Stop new material. Re-read cheat sheets and final-review.md only.
Exam day:                Read final-review.md once over coffee. Eat. Go pass it.

Time-budget per resource (single sitting, end-to-end)

Resource Realistic time
Each numbered topic file 30–45 min reading + 15 min experimentation
Each section index 5 min
One cheat sheet 15–20 min
One domain's practice questions (15–20 Qs) 30–45 min
Mock Exam (45 Qs, timed) 90 min + 30 min review
final-review.md 20 min

Roadmap for the guide itself

This guide ships as a living resource. The roadmap below is what's planned for the next two quarters — issues and PRs against any of these are welcome.

Q2–Q3 2026 (next 6 months)

  • ✅ Align to March 2026 blueprint — complete
  • ✅ MIT license + open-source release — complete
  • ✅ 2026 update callouts in overview and final-review — complete
  • ✅ Mock-exam debrief files mapping every question to a topic file — complete
  • ✅ End-to-end RAG worked example in code-examples/tsql/ — complete
  • ✅ Mermaid diagrams in Domain 3 topic files — complete
  • ✅ Case-study mini-blocks in both mock exams (mirroring real DP-800 format) — complete
  • ✅ Practice-question rebalancing (+2 Hard Domain 1, +2 Easy Domain 2, +1 REGEXP) — complete
  • ✅ Mental-model phrasings in highest-leverage Domain 1 / 2 topics — complete
  • ✅ Hands-on lab pack — runnable T-SQL scripts walking through vector search, RAG, full-text, and MCP scenarios — see certification/resources/labs/labs.md (4 labs · ~1,870 lines)
  • ⏸ Half-precision vector examples — feature still public preview, no published examples yet; deferred until GA
  • ✅ Microsoft Foundry walkthrough as an embedding-maintenance method — see certification/09-models-embeddings/02-embedding-maintenance.md#method-7-microsoft-foundry
  • 🔄 Video walkthroughs of the hardest topics — outside the text-only scope of this guide; community contributions welcome

Q4 2026 (3–6 months out)

  • ⏳ Azure SQL DiskANN GA content updates — currently public preview on SS2025 + Azure SQL + Fabric; will refresh on GA announcement (last verified 2026-05-22 — learn.microsoft.com/en-us/sql/sql-server/ai/vectors: "Approximate vector index and vector search are in preview")
  • ⏳ Half-precision vector GA content updates — currently preview; will refresh on GA (last verified 2026-05-22 — learn.microsoft.com/en-us/sql/t-sql/data-types/vector-data-type: "float16 vector is currently available for preview")
  • ⏳ Updated mock exams following any Microsoft blueprint refresh — blueprint verified unchanged as of 2026-05-22 (still "Skills measured as of March 12, 2026", page updated_at: 2026-03-23 is a cosmetic edit); will refresh when Microsoft bumps the "Skills measured as of" date
  • ⏳ Community contributor list in CONTRIBUTORS.md — awaiting first community PR (last verified 2026-05-22 — all 11 merged PRs to date are from @kengio)
  • ✅ Spaced-repetition deck (Anki) generated from cheat-sheet facts — see certification/resources/anki/ (130 cards across 6 cheat sheets)
  • ✅ Translation scaffolding so non-English learners can fork and translate — see TRANSLATING.md

Q1 2027 (6–12 months out)

Legend: ✅ done · 🔄 in progress / next up · ⏸ deferred until upstream GA · ⏳ planned · 🌱 ideas being explored

Quick navigation

Resource Description
Start Studying → Main index with all 11 study sections and progress tracker
Cheat Sheets Seven quick-reference guides for exam day
Practice Questions 60+ domain-specific questions with explanations
Mock Exam 1 50-question timed practice exam (45 standalone + 5-question case study)
Mock Exam 1 Debrief Per-question map to topic files + cheat sheets, plus a study plan by miss count
Mock Exam 2 Second 50-question practice exam (different questions; includes case study)
Mock Exam 2 Debrief Same debrief pattern for Mock 2
RAG Walkthrough End-to-end RAG pipeline in ~80 lines of T-SQL
Final Review 20-minute exam-morning scan
Exam Tips Time management, common traps, and strategy
Appendix Glossary, comparison tables, error reference
T-SQL Code Examples Standalone runnable T-SQL snippets

Repository layout

dp-800-study-guide/
├── certification/
│   ├── dp-800-overview.md           # main entry point — start here
│   ├── 01-database-objects/         # tables, indexes, JSON, partitioning
│   ├── 02-programmability-objects/  # views, functions, procedures, triggers
│   ├── 03-advanced-tsql/            # CTEs, window functions, regex, graph
│   ├── 04-ai-assisted-tools/        # GitHub Copilot, MCP server endpoints
│   ├── 05-data-security-compliance/ # encryption, RLS, DDM, secure endpoints
│   ├── 06-performance-optimization/ # configs, isolation, plans, DMVs
│   ├── 07-cicd-database-projects/   # SQL DB Projects, schema drift
│   ├── 08-azure-services-integration/ # DAB, REST/GraphQL, CDC/CT/CES
│   ├── 09-models-embeddings/        # external models, embedding maintenance
│   ├── 10-intelligent-search/       # full-text, vector, hybrid (RRF)
│   ├── 11-rag/                      # RAG, sp_invoke_external_rest_endpoint
│   └── resources/
│       ├── cheat-sheets/            # quick-reference for exam day
│       ├── practice-questions/      # per-domain Q&A
│       ├── mock-exam/               # mock exam 1
│       ├── mock-exam-2/             # mock exam 2
│       ├── code-examples/tsql/      # standalone T-SQL examples
│       ├── appendix/                # glossary, comparisons, error reference
│       ├── final-review.md          # 20-minute exam-morning scan
│       ├── exam-tips.md             # strategy and time management
│       ├── anki/                    # spaced-repetition deck (130 cards) + import readme
│       └── official-links.md        # Microsoft docs and exam registration
├── practice/                        # adaptive practice quiz — HTML/JS/CSS + build.py + JSON banks
│                                    # auto-deployed to GitHub Pages
├── i18n/                            # community translations — parallel tree per locale
├── .github/workflows/               # CI: markdownlint + lychee (lint.yml), Pages deploy (deploy-practice.yml)
├── CHANGELOG.md                     # versioned change log
├── CONTRIBUTING.md / CONTRIBUTORS.md # contribution guide and roster
├── TRANSLATING.md                   # translation conventions
├── LICENSE                          # MIT
└── README.md                        # this file

Official Microsoft resources

📋 Exam and certification

Exam and certification

📚 Documentation by topic

Documentation by topic

👥 Community and learning paths

Community and learning paths

Translations

The English content under certification/ is canonical. Community translations live in i18n/<locale>/ as parallel trees and don't alter the English source.

  • Available locales — see i18n/README.md (none yet — be the first!)
  • How to translate — read TRANSLATING.md for BCP-47 locale codes, layout, priority order, and the currency policy
  • Coordinate first — open an issue titled i18n: <locale name> so two people don't start the same locale in parallel

Contributing

Found an error, a stale link, or a topic that needs deeper coverage? PRs are welcome.

  • Small fixes (typos, link rot, factual corrections) — open a PR directly
  • New practice questions or topic expansions — open an issue first to discuss scope
  • Blueprint changes — Microsoft updates DP-800 periodically; PRs that bring sections in line with the latest skills-measured list are especially appreciated

Please keep the existing structure: each topic file follows the conventions in CLAUDE.md, and code examples live in certification/resources/code-examples/tsql/.

License

Released under the MIT License. Use, fork, remix, redistribute — just keep the copyright notice.

This guide is a community resource. It is not affiliated with, endorsed by, or sponsored by Microsoft.
"Microsoft", "Azure", "SQL Server", and "Microsoft Fabric" are trademarks of Microsoft Corporation.
Always verify against the official DP-800 skills measured page — it is the source of truth.

Good luck on the exam. You've got this. ⭐ this repo if it helped you pass.

About

Open-source community study guide for Microsoft DP-800: Developing AI-Enabled Database Solutions. Aligned to the March 2026 official blueprint. Covers SQL Server 2025, Azure SQL, Microsoft Fabric, vector search, RAG, MCP, and more.

Topics

microsoft sql-server ai mcp embeddings certification study-guide obsidian tsql rag exam-prep vector-search azure-sql diskann microsoft-fabric dp-800

Resources

Readme

License

MIT license

Contributing

Contributing
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases 3

v1.2.0 — Hands-on labs + renewal + companion exams + Foundry walkthrough Latest
May 21, 2026
+ 2 releases

Packages

 
 
 

Contributors

Languages