Home › Certifications › DP-700 Certification Guide
Microsoft Certification Guide · 2026DP-700 Certification Guide 2026: Fabric Data Engineer Associate
Everything about the DP-700 exam — all three domains and weightings, the three query languages it demands, why a full third of it happens after your pipeline is already in production, all 54 examinable objectives mapped, data engineer salary bands, and an 8-week roadmap. This is the exam that replaced the retired DP-203.
Authorized Training & Technology Partners
Microsoft
DP-700 — Fabric Data Engineer Associate
The credential this guide covers
Claude
Claude as your data engineering copilot
Built into the same programme — read the CCAR-F guideDP-700 (Implementing Data Engineering Solutions Using Microsoft Fabric) is the certification that replaced the retired DP-203 as Microsoft’s data engineering credential. Three domains, all weighted 30–35%, pass at 700 out of 1000. The two things that catch candidates out: it assumes fluency in three query languages, and roughly a third of the exam is about what happens after your solution is already running.
Microsoft states the language requirement outright in the audience profile: “You should be skilled at manipulating and transforming data by using Structured Query Language (SQL), PySpark, and Kusto Query Language (KQL).” Most candidates arrive with SQL. Many have PySpark. Very few have KQL — and it appears across the blueprint, not in one corner of it.
DP-700 exam at a glance
| Attribute | Detail |
|---|---|
| Exam code | DP-700 — Implementing Data Engineering Solutions Using Microsoft Fabric |
| Certification earned | Microsoft Certified: Fabric Data Engineer Associate |
| Level | Associate — the role-based working tier |
| Passing score | 700 out of 1000 |
| Replaces | DP-203 (Azure Data Engineer Associate), now retired |
| Domains | Three — each weighted 30–35% |
| Examinable objectives | 54 across 10 sub-domains |
| Languages assumed | SQL, PySpark and KQL — stated in the audience profile |
| Skills measured version | Current as of 21 July 2026 |
| Delivery | Pearson VUE — online proctored or at a test centre |
| Renewal | Annually — associate certifications expire each year, renewed by a free online assessment on Microsoft Learn |
| Practice assessment | A free official practice assessment is published on Microsoft Learn |
What is the DP-700 certification?
Microsoft’s audience profile asks for subject matter expertise with data loading patterns, data architectures, and orchestration processes, and names three responsibilities: ingesting and transforming data, securing and managing an analytics solution, and monitoring and optimizing an analytics solution.
Two of those three are operational. That is the shape of this credential — it certifies someone who can build a pipeline and keep it alive, which is a materially harder thing than building one.
Kusto Query Language is the gap that sinks otherwise strong candidates. It is not optional garnish: the blueprint names it in transform data by using PySpark, SQL, and KQL, in choose between Dataflows Gen2, notebooks, KQL, and T-SQL, and again in process data by using KQL inside the streaming sub-domain. If you have never written KQL, budget real weeks for it — not an afternoon.
DP-700 skills measured and weightings
Three domains, and Microsoft has weighted them as close to identically as a published range allows.
DP-700 exam blueprint — share of exam by domain
Official Microsoft weightings, skills measured as of 21 July 2026. All three domains share an identical 32.5% midpoint.
Source: Microsoft Learn, “Study Guide for Exam DP-700”, skills measured as of 21 July 2026. Three equal domains means the monitoring and optimisation work most engineers treat as an afterthought is worth exactly as much as the ingestion work they enjoy.
Equal thirds is a deliberate signal. Monitoring and optimisation — 30–35% — is entirely post-deployment work: resolving pipeline, Dataflow, notebook, Eventhouse, Eventstream, T-SQL and OneLake shortcut errors, then optimising Lakehouse tables, pipelines, warehouses, Eventstreams, Spark and queries. You cannot pass this exam by being good at building things.
Where the 54 objectives actually cluster
Domain weighting tells you how the marks are split. Counting the published objectives tells you where the detail is — and those are not the same thing.
Examinable objectives by sub-domain — the six densest
Count of published bullet-level objectives in each sub-domain of the DP-700 blueprint
Bar length is relative to the largest sub-domain. Nothing is hidden: the four sub-domains not charted are Configure Fabric workspace settings (4), Implement lifecycle management (3), Orchestrate processes (3) and Design and implement loading patterns (3) — 13 objectives. The six shown carry 41. Total 54. Counted from the published blueprint, skills measured as of 21 July 2026.
Three of the six densest sub-domains sit in Monitor and optimize. Together, error resolution, performance optimisation and monitoring account for 17 of the 54 objectives — more than either of the other two domains contributes on its own. The exam is telling you exactly where it will push, and it is not where most study plans go.
The three domains as flashcards
Implement and manage an analytics solution
Workspace settings: Spark, domain, OneLake and Apache Airflow workspace settings. Lifecycle management: version control, database projects, deployment pipelines. Security and governance (the densest sub-domain in this domain, 8 objectives): workspace-level, item-level and row/column/object/folder-file-level access controls, dynamic data masking, sensitivity labels, endorsing items, Fabric audit logs, OneLake security. Orchestration: choosing between Dataflow Gen2, a pipeline and a notebook; schedules and event-based triggers; orchestration patterns with parameters and dynamic expressions.
30–35% · 18 objectivesIngest and transform data
Loading patterns: full and incremental loads, preparing data for a dimensional model, a loading pattern for streaming data. Batch: choosing a data store, choosing between Dataflows Gen2 / notebooks / KQL / T-SQL, OneLake shortcuts, mirroring, pipeline ingestion, transforming with PySpark, SQL and KQL, denormalising, grouping and aggregating, and handling duplicate, missing and late-arriving data. Streaming: choosing a streaming engine, native tables versus OneLake shortcuts in Real-Time Intelligence, query acceleration, Eventstreams, Spark structured streaming, KQL, and windowing functions.
30–35% · 19 objectivesMonitor and optimize an analytics solution
Monitor: data ingestion, data transformation, semantic model refresh, configuring alerts. Resolve errors in seven distinct places: pipelines, Dataflow Gen2, notebooks, Eventhouse, Eventstream, T-SQL and OneLake shortcuts. Optimize: a Lakehouse table, a pipeline, a data warehouse, Eventstreams and Eventhouses, Spark performance, and query performance.
30–35% · 17 objectivesThree languages, not one
SQL / T-SQL for warehouse work and error resolution. PySpark for notebook transformation and structured streaming. KQL for Real-Time Intelligence, Eventhouse queries and streaming processing. The audience profile names all three by name. A candidate fluent in only SQL is prepared for perhaps a third of the transformation objectives.
Stated in the audience profileReal-Time Intelligence is not a footnote
Eventstreams, Eventhouse, native tables versus shortcuts, query acceleration, Spark structured streaming, windowing functions — streaming carries 7 objectives in ingestion alone, plus Eventhouse and Eventstream error resolution and optimisation in domain 3. If your Fabric experience is batch-only, this is your largest blind spot after KQL.
Across domains 2 and 3Coming from DP-203?
DP-700 is its replacement, but it is not a rename. DP-203 was an Azure exam — Synapse, Data Factory, Databricks. DP-700 is a Fabric exam: OneLake, Lakehouse, Eventhouse, Dataflow Gen2, deployment pipelines. The data engineering concepts transfer; almost none of the specific product knowledge does.
Migration noteWho DP-700 is actually for
- ETL and SQL developers moving from stored procedures and SSIS into a lakehouse architecture.
- DP-203 holders whose credential has retired and who need the Fabric-native successor.
- Power BI analysts moving upstream — the natural next step after PL-300, and the one that changes your salary band rather than your job title.
- Azure data engineers whose organisation is consolidating Synapse and Data Factory workloads onto Fabric.
Where it will not carry you
- It is not a beginner credential. “Subject matter expertise” is Microsoft’s own phrasing. Start at DP-900 if you are new to data.
- It is not a machine learning credential. You will move and shape data for models; you will not build them.
- It is Fabric-specific. The architectural thinking is portable; the product surface is not. Outside a Fabric shop, this certifies judgement more than tooling.
The data engineering salary premium
This is the credential in our Microsoft catalogue with the clearest pay story. Data engineering sits structurally above data analysis in every market we looked at — because pipelines that break cost money in a way that a stale dashboard does not.
How that premium is built
Entry
Mid
Senior
Lead
Data engineer pay by market
India and the US shown as two separate charts, because rupee and dollar bands are different measures on different scales and should never share an axis.
India — annual CTC
Data engineer roles, ₹ lakh per annum
Bar length maps the upper bound of each band against a ₹50 L scale. Bands compiled from published Indian data engineer ranges. Streaming and real-time experience sits at the top of each band; batch-only ETL work at the bottom.
United States — annual base
Data engineer roles, US$ thousands
Bar length maps each figure against a $190K scale. Bands compiled from published US data engineer ranges; base pay only, excluding bonus and equity.
The ~35% figure is ours, not a survey’s. We took the midpoint of each Indian band above and compared it with the midpoint of the equivalent band on our PL-300 page: entry ₹7 L vs ₹5 L, mid ₹14 L vs ₹10 L, senior ₹24 L vs ₹18 L, lead ₹37.5 L vs ₹29.5 L. That is +40%, +40%, +33%, +27% — averaging roughly 35%. It is a modelled comparison of two published role bands, not a measured outcome of holding DP-700, and your result will depend on your market and employer.
Compensation figures are compiled from independent, publicly available industry sources and are shown for role context. They are not a guarantee of pay in any specific market, company or outcome, and 360DT does not promise a salary result from any certification or programme.
Why DP-700 matters right now
1. DP-203 retired, and the replacement is not a rename
A large population of certified Azure data engineers holds a credential that no longer exists. DP-700 is the successor, but it moved the ground: Synapse and Data Factory gave way to OneLake, Lakehouse, Eventhouse and Dataflow Gen2. That creates a genuine re-certification market, and it means a current DP-700 distinguishes you from a lapsed DP-203 in a way that certifications rarely do.
2. Fabric consolidated a stack that used to be five products
The blueprint reads like an inventory of what Fabric absorbed: pipelines, notebooks, warehouses, lakehouses, real-time streams, semantic models and deployment tooling, all with one security model and one storage layer in OneLake. Organisations migrating onto that platform need people who understand the whole surface, not one corner of it.
3. Streaming has stopped being a specialism
Eventstreams, Eventhouse, structured streaming and windowing functions sit in the core ingestion domain, not in an optional module. Real-time intelligence has moved from something a few teams did to something a data engineer is simply expected to handle — and the exam reflects that shift.
4. Operations is where the marks and the money are
Seven distinct error-resolution objectives and six optimisation objectives tell you what employers actually pay for. Anyone can stand up a pipeline in a demo tenant. Diagnosing why an Eventstream is dropping records at 3am, or why a Lakehouse table is scanning far more than it should, is the skill that separates pay bands.
Your 8-week DP-700 roadmap
Weighted to three equal domains, with KQL deliberately started early because it is the language most candidates lack. Assume 8–10 hours per week — this is an Associate exam that assumes existing expertise.
Fabric foundations and workspace configuration
OneLake, Lakehouse, Warehouse, Eventhouse — what each is for and when to choose it. Configure Spark, domain, OneLake and Apache Airflow workspace settings. Start KQL this week, thirty minutes a day, and keep it running for the whole eight weeks.
Security, governance and lifecycle
The densest sub-domain in domain 1. Workspace-level, item-level, and row, column, object and folder/file-level access controls. Dynamic data masking. Sensitivity labels. Endorsing items. Fabric audit logs. OneLake security. Then version control, database projects and deployment pipelines.
Orchestration
The decision that recurs throughout the exam: Dataflow Gen2 versus a pipeline versus a notebook, and why. Schedules and event-based triggers. Orchestration patterns with notebooks and pipelines, including parameters and dynamic expressions. Completes domain 1.
Loading patterns and batch ingestion
Full versus incremental loads. Preparing data for a dimensional model. Choosing a data store. OneLake shortcuts. Mirroring. Pipeline ingestion. Denormalising, grouping, aggregating. Handling duplicate, missing and late-arriving data — the objective that most resembles a real incident.
Transformation in all three languages
The same transformation implemented in PySpark, SQL and KQL, deliberately, side by side. Then the choice objective: Dataflows Gen2 versus notebooks versus KQL versus T-SQL, and the trade-offs that decide it. This week is the one that most affects your score.
Streaming and Real-Time Intelligence
Choosing a streaming engine. Native tables versus OneLake shortcuts in Real-Time Intelligence, and query acceleration for shortcuts. Eventstreams. Spark structured streaming. Processing with KQL. Windowing functions. Completes domain 2.
Monitoring and error resolution
Monitor ingestion, transformation and semantic model refresh. Configure alerts. Then break things on purpose and fix them: pipeline, Dataflow Gen2, notebook, Eventhouse, Eventstream, T-SQL and OneLake shortcut errors. Seven error types, seven deliberate failures.
Optimisation, practice assessment, book the exam
Optimise a Lakehouse table, a pipeline, a data warehouse, Eventstreams and Eventhouses, Spark performance and query performance. Completes domain 3. Then the free official practice assessment — score it by domain, not overall, because three equal domains means one weak third is a fail.
Register with a personal MSA account, not a work or school account. Exam records tied to an organisational account are unrecoverable if you leave — and with an annually renewing credential you will need that record every year.
Learn from Microsoft Certified Trainers
Certification prep at 360DT is delivered live by Microsoft Certified Trainers, Anthropic Authorized Instructors and a Microsoft Business Applications MVP — our full international faculty.
Shantanu Pandey
Microsoft, Google, NVIDIA and HPE Certified Trainer
Director of Engineering at Meteoros Automation, ranked in Microsoft's Top 100 Trainers of 2025, holding MCT, Google Cloud, NVIDIA and PeopleCert credentials, with 500+ trainings delivered since 2011.
Sid
Anthropic Authorized Instructor and Azure Data Engineer
Dynamic corporate trainer and consultant with more than 10 years of experience across Microsoft Fabric, Azure Data Engineering, and modern cloud platforms. Focused on clear, practical learning that sticks.
Arshad Ahmad
Microsoft Certified Trainer (MCT), Cybersecurity & Power Platform Expert
Technology Trainer with 15 years in strategic client delivery, holding MCT, Azure Solutions Architect Expert, Cybersecurity Architect Expert and Power BI Data Analyst credentials, with 200+ trainings delivered.
Sravia
NVIDIA Authorized Instructor and AI Data Specialist
Passionate AI and data evangelist with over 10 years of experience across machine learning, MLOps, and cloud native technologies. Dedicated to delivering impactful, engaging enterprise training programs.
Johan
Microsoft Business Applications MVP and Microsoft Certified Trainer
D365 Customer Service Lead – Europe at Avanade, and founder of the Power Platform School. A Microsoft Business Applications MVP and Microsoft Certified Trainer based in London, specializing in Dynamics 365 Customer Service and Power Platform.
Vikas Mittal
Google and Anthropic Authorized AI Instructor
Veteran technology trainer with more than 25 years of rich industry experience spanning Google, Microsoft, and emerging AI platforms. An exceptional mentor recognized for proven teaching excellence.
Deep
Anthropic Authorized Instructor and AI Cloud Specialist
Seasoned technology leader with over 20 years of experience across AI, cloud computing, and enterprise training. Known for simplifying complex ideas and delivering practical, real world learning.
Bipeen
AWS, Microsoft, and Anthropic Certified Trainer
AI, machine learning, and cloud transformation expert with more than 25 years of experience. A renowned global corporate trainer and conference speaker known for practical, outcome driven sessions.
Akim Nyamande
Microsoft, CompTIA, Juniper and Cisco Certified Trainer
IT training facilitator with over five years of experience and certifications across Microsoft, CompTIA, Juniper, and Cisco. Previously a Network Administrator before moving into technical training, now delivering hands-on courses in networking, systems administration, and cybersecurity fundamentals.
Ali El Khatib
Cisco Certified Systems Instructor (CCSI) and Microsoft Certified Trainer
Infrastructure Engineer at RHUH with 13 years in the training field, holding CCSI, MCSE, CCNP Routing and Switching, CCNP Security, CompTIA and Microsoft Azure certifications.
Florian Garcia Compte
Cisco Certified Systems Instructor (CCSI)
Partner Director at AVAANZA FORMACION, a Cisco Learning Partner based in Madrid. Certified Cisco Systems Instructor (CCSI No. 21053) specializing in Cisco Data Center (ACI/SDN), Nexus, and Wireless, delivering official Cisco certification courses including CCNA, CCNP Enterprise, and CCNP Data Center.
Data Engineer Program — DP-700 + DP-900
The 360DT Data Engineer Program covers the full DP-700 blueprint — OneLake and Lakehouse architecture, batch and streaming ingestion, PySpark, SQL and KQL, security and governance, and the monitoring and optimisation work that is a third of the exam — plus DP-900 fundamentals and Claude as your data engineering copilot. Replaces the retired DP-203 track.
Explore the Data Engineer Program
DP-700 frequently asked questions
Is DP-700 the replacement for DP-203?
Yes. DP-700 is Microsoft’s current data engineering certification and DP-203 has retired. But it is not a rename — DP-203 examined Azure Synapse, Data Factory and Databricks, while DP-700 examines Microsoft Fabric: OneLake, Lakehouse, Eventhouse, Dataflow Gen2 and deployment pipelines. Your data engineering concepts transfer; most of the product-specific knowledge does not.
Do I really need KQL for DP-700?
Yes. The audience profile states you should be skilled at manipulating and transforming data using SQL, PySpark and Kusto Query Language. KQL appears in transformation objectives, in the choice between Dataflows Gen2, notebooks, KQL and T-SQL, and in streaming processing. It is the single most common preparation gap, and it needs weeks rather than an afternoon.
What is the hardest part of the DP-700 exam?
Two things. KQL, for candidates who have never written it. And the monitor-and-optimize domain, which is 30–35% of the exam and covers seven distinct error types plus six optimisation targets — work that only makes sense if you have actually operated a pipeline rather than only built one.
How many objectives are on the DP-700 blueprint?
54 published bullet-level objectives across 10 sub-domains, as of the 21 July 2026 skills-measured version. They are distributed 18 / 19 / 17 across the three domains, which corroborates the equal 30–35% weighting. The densest single sub-domain is batch ingestion and transformation, with 9.
How long does it take to prepare for DP-700?
Around eight weeks at 8–10 hours per week for a working data or ETL engineer — the roadmap on this page. Add several weeks if KQL is new to you, and add more if you have never worked with streaming data, since Real-Time Intelligence carries seven ingestion objectives on its own.
Should I take DP-900 before DP-700?
If you are new to data, yes — DP-900 (Azure Data Fundamentals) establishes the vocabulary DP-700 assumes you already have. If you are a working SQL or ETL developer, you can go straight to DP-700. Our programme covers both, because the fundamentals cost little time and close gaps people do not know they have.
PL-300 or DP-700 — which is the better move?
Different jobs. PL-300 is the presentation and modelling layer and is the faster route into a data role from a business background. DP-700 is upstream engineering and pays materially more — roughly 35% more at equivalent experience on the bands shown above — but it assumes real technical depth. Analysts commonly do PL-300 first and DP-700 second.
Does the DP-700 certification expire?
Yes. Microsoft associate certifications expire annually, and DP-700 is Associate tier. Renewal is free through a short online assessment on Microsoft Learn during your renewal window — considerably less work than the original exam, but you do have to remember to do it.
Sources and further reading
- Microsoft Learn — Study Guide for Exam DP-700: Implementing Data Engineering Solutions Using Microsoft Fabric (audience profile naming SQL, PySpark and KQL; three skills-measured domains at 30–35% each; full objective list; skills measured as of 21 July 2026; change log)
- Microsoft Learn — exam scoring and score reports (a score of 700 or greater is required to pass)
- Microsoft Learn — certification renewal (associate, expert and specialty certifications expire annually and renew via a free online assessment)
- Objective counts on this page were tallied directly from the published blueprint: 18 / 19 / 17 across the three domains, 54 in total
- Published Indian and US data engineer compensation ranges, used for role context only; the ~35% premium is our own comparison against the analyst bands on our PL-300 guide, not a survey figure
360DT is an independent training provider. Microsoft certification exams are administered by Microsoft through Pearson VUE and are not included in programme tuition. Exam details are accurate as of 20 August 2026; always confirm current format, pricing, renewal policy and skills measured on Microsoft Learn before booking.