CareerGuide 16 of 36
Career guide · Guide 16 of 36
DP-700 Fabric Data Engineer Associate: Exam Guide and Study Plan
Microsoft DP-700 Fabric Data Engineer exam guide: format, syllabus and the 3 skill areas, how it replaced DP-203, a 6-week study plan and practice questions.
On this page
DP-700 is the exam for the Microsoft Certified: Fabric Data Engineer Associate certification. It checks that you can build data engineering solutions in Microsoft Fabric: load and transform data, run and secure it, and keep it fast and healthy. It replaced the older Azure data engineer exam, DP-203, which is why searches for DP-203 are falling while DP-700 is rising.
This guide explains the exam in plain words, what Fabric is, what to study, and a realistic plan to pass.
The exam at a glance
These facts come from the official Microsoft Learn pages, checked on 31 October 2026.
| Certification | Microsoft Certified: Fabric Data Engineer Associate |
| Time | 100 minutes |
| Passing score | 700 (on Microsoft’s scale of 1–1,000) |
| Price | Depends on the country where you take it |
| Where | Proctored, booked through Pearson VUE |
| Languages | English, Japanese, Chinese (Simplified and Traditional), German, French, Italian, Korean, Spanish, Portuguese (Brazil) |
| Languages you must know | SQL, PySpark and KQL (Kusto Query Language) |
| Retake | 24 hours after a first failed attempt |
| Renewal | Free online assessment on Microsoft Learn before it expires |
The exam may include interactive questions, not only multiple choice. Microsoft also offers a free practice assessment and an exam sandbox, so you can try the question types first.
First, what is Microsoft Fabric?
Fabric is Microsoft’s all-in-one analytics platform. Everything sits on OneLake, a single data lake for the whole organisation, and tables are stored in the open Delta Lake format. On top of OneLake you use:
- Lakehouses: files and Delta tables, worked on with Spark notebooks (PySpark or Spark SQL).
- Warehouses: a SQL data warehouse you work on with T-SQL.
- Pipelines and Dataflows: data movement and low-code transformation, similar to Azure Data Factory.
- Real-Time Intelligence: eventstreams to capture streaming data and eventhouses (KQL databases) to query it fast with KQL.
- Shortcuts to point at data in other places (such as ADLS or S3) without copying it, and mirroring to keep a copy of an operational database in sync.
If you know Spark, Delta Lake and SQL, most of Fabric will feel familiar. KQL and the Fabric-specific features are usually what people need to learn.
What the exam covers
The exam has three skill areas, each worth 30–35% of the questions.
| Skill area | Weight | In plain words |
|---|---|---|
| Implement and manage an analytics solution | 30–35% | Workspace settings, version control and deployment pipelines, security (workspace, item, row, column and file level, masking, sensitivity labels), and orchestration with pipelines, notebooks, schedules and triggers. |
| Ingest and transform data | 30–35% | Full and incremental loads, preparing data for a dimensional model, batch work with pipelines, dataflows, notebooks and T-SQL, shortcuts and mirroring, and streaming with eventstreams, Spark Structured Streaming and KQL, including windowing. Handling duplicate, missing and late data. |
| Monitor and optimize an analytics solution | 30–35% | Monitoring loads, transformations and semantic model refreshes, setting alerts, fixing errors in pipelines, dataflows, notebooks, eventstreams, eventhouses and T-SQL, and tuning lakehouse tables, warehouses, Spark and queries. |
Because the three areas are weighted almost equally, you cannot skip any of them.
What to study, with lessons on this site
DataDank does not have a Fabric course, but most of the core skills are the same everywhere. Use these lessons for the general skills, and Microsoft Learn for the Fabric-specific features and KQL.
| Skill | Study on this site |
|---|---|
| PySpark transformations | PySpark course, especially window functions and nested JSON |
Delta tables and table maintenance (OPTIMIZE, file sizes) |
Delta Lake course, OPTIMIZE and liquid clustering |
| Medallion layers in a lakehouse | Medallion architecture |
| Full vs incremental loads | Incremental loading, watermarks and backfills |
| Dimensional models | Star schema and slowly changing dimensions |
| Streaming and windows | Spark Structured Streaming |
| Late and duplicate data | Idempotency in pipelines |
| SQL performance | Query optimisation fundamentals |
A 6-week study plan
This plan assumes about 1 hour a day. Use a Fabric trial capacity so you can try every feature; the exam rewards people who have clicked through the real screens.
| Week | Focus | Do this |
|---|---|---|
| 1 | Fabric basics | Create a workspace, a lakehouse and a warehouse; load a CSV; explore OneLake and create a shortcut. |
| 2 | Batch ingestion | Build a pipeline with a Copy activity; do a full and an incremental load; transform with a notebook and with a dataflow. |
| 3 | Modelling and T-SQL | Build a small star schema in the warehouse; practise T-SQL; handle duplicates and late-arriving rows. |
| 4 | Real-time | Create an eventstream into an eventhouse; learn basic KQL (where, summarize, bin() for time windows). |
| 5 | Manage and secure | Set up Git integration and a deployment pipeline; apply row-level security, column-level security and masking. |
| 6 | Monitor, optimise, review | Use the Monitoring hub; fix a failed pipeline; tune a lakehouse table; take the official practice assessment. |
Practice questions
These are our own questions, written to show the style and the reasoning. They are not taken from the real exam.
Question 1
Your data already lives in an ADLS Gen2 account. You want it to appear in a Fabric lakehouse without copying it. What should you use?
Show answer
Answer: B. A shortcut points to data in another location, so it appears in OneLake without being copied. A Copy activity (A) and a dataflow (D) would copy it.
Question 2
You need a near real-time copy of an operational Azure SQL Database in OneLake, kept in sync automatically, with no pipelines to build. What should you use?
Show answer
Answer: B. Mirroring continuously replicates a supported source database into OneLake. Shortcuts (A) work for storage locations, not for keeping a database in sync.
Question 3
IoT devices send thousands of readings per second, and engineers want to query the last hour with KQL. Which combination fits best?
Show answer
Answer: B. Eventstreams capture streaming data, and eventhouses (KQL databases) are built for fast queries over large volumes of time-based data with KQL.
Question 4
You want to move workspace content from development to test to production in a controlled way. What should you use?
Show answer
Answer: B. Deployment pipelines promote Fabric items between stages, which is part of lifecycle management, together with Git version control.
Exam-day tips
- Learn the “choose between” objectives. Several skills are literally about choosing: pipeline or notebook, dataflow or notebook or T-SQL, which streaming engine, which data store. Know the strengths of each.
- Practise KQL basics. People coming from SQL often lose easy marks on KQL syntax.
- Expect preview features. Microsoft says most questions cover generally available features, but commonly used preview features can appear too.
- Use the free practice assessment and the exam sandbox before the real exam.
A word on “dumps”
Sites that sell “real exam questions” break the Microsoft exam agreement and can lead to your certification being revoked. Fabric also changes quickly, so those question lists are often out of date. Learn it hands-on with a trial capacity and use the official practice assessment.
Is it worth it?
DP-700 is the current Microsoft data engineering certification, so it is the one to take if you work with Azure or Microsoft Fabric, or want to. If you hold DP-203, which has been retired, DP-700 is the natural next step. As with any certificate, pair it with a small project you can explain in an interview.