What are the current DP-700 domain weights?

Implement and manage an analytics solution is 30-35%, ingest and transform data is 30-35%, and monitor and optimize an analytics solution is 30-35%.

Which languages should a DP-700 candidate practice?

The official audience profile expects SQL, PySpark, and Kusto Query Language (KQL). Practice each in the appropriate Fabric engine.

Does DP-700 cover both batch and streaming data engineering?

Yes. Objectives include full and incremental batch loading, pipelines, Dataflow Gen2, notebooks, shortcuts and mirroring, plus Eventstreams, Spark Structured Streaming, KQL, Eventhouse, windowing, and streaming-engine selection.

Are PrepKloud DP-700 questions copied from the exam?

No. PrepKloud creates original scenarios from public objectives and official Microsoft documentation and does not use dumps, leaks, or recalled live questions.

DP-700 Microsoft Fabric Data Engineer Roadmap

Q: Is DP-700 active in August 2026?

Yes. Exam DP-700 and the Fabric Data Engineer Associate certification are active as of August 19, 2026. The current English skills measured took effect July 21, 2026; verify the official guide before scheduling.

Prepare for Implementing Data Engineering Solutions Using Microsoft Fabric through five connected phases: governed workspace foundations, batch engineering, real-time engineering, lifecycle and security, then monitoring, optimization, projects, and exam readiness.

Exam: DP-700Fabric Data Engineer AssociateFive phasesSQL · PySpark · KQL

Current scope: DP-700 is active as of August 19, 2026. The current English skills measured took effect July 21, 2026. Microsoft can revise objectives, features, and localized exams, so verify the official DP-700 study guide before scheduling. PrepKloud is independent and uses original scenarios grounded in public Microsoft documentation—never dumps, leaks, or recalled live questions.

Official domain balance

The current blueprint is deliberately balanced. Study the domains together: an incremental pipeline is incomplete if reruns duplicate data, its workspace exposes raw PII, or a morning capacity peak prevents the gold table from meeting its freshness objective.

30-35% · Implement and manage an analytics solutionSpark, domain, OneLake, and Airflow settings; version control, database projects, deployment pipelines; granular security, masking, labels, endorsement, audit, OneLake security; Dataflow Gen2, pipeline, notebook, schedules, event triggers, parameters, and expressions.

30-35% · Ingest and transform dataFull, incremental, dimensional, and streaming loading patterns; store/engine selection; shortcuts, mirroring, pipelines, Dataflow Gen2, PySpark, SQL, KQL, quality, Eventstreams, Structured Streaming, native tables, acceleration, and windows.

30-35% · Monitor and optimize an analytics solutionIngestion, transformation, semantic refresh, alerts, pipeline/Dataflow/notebook/Eventhouse/Eventstream/T-SQL/shortcut diagnosis, and optimization across lakehouse, Warehouse, Spark, KQL, queries, streams, and capacity.

Fabric foundations, storage, and workspace control

Weeks 1-2

Begin with architecture and effective access. Learn why OneLake is shared storage, where workspaces and domains fit, and how each engine changes language, latency, and operations.

Explain OneLake, workspaces, domains, items, capacities, environments, and Spark workspace defaults.

Choose Lakehouse for open Delta/Spark engineering, Warehouse for SQL-first relational analytics, and Eventhouse for high-volume time-series and KQL.

Compare copy, Dataflow Gen2, pipeline, notebook, mirroring, OneLake shortcut, Eventstream, and Structured Streaming paths.

Configure domain, OneLake, Spark, and Apache Airflow workspace settings from governance and workload requirements.

Map workspace roles, item rights, OneLake security roles, SQL/KQL access, shortcut behavior, and compute permissions for each persona.

Create measurable contracts for grain, schema, quality, freshness, lineage, retention, cost, replay, and ownership.

Batch ingestion, medallion transformation, and orchestration

Weeks 3-5

Build complete and incremental paths that can restart safely. Use the right low-code, orchestration, SQL, and Spark surfaces instead of forcing one tool to perform every task.

Design full bootstrap and incremental loads with source boundaries, per-entity watermarks, idempotent merge, deletion behavior, and backfill.

Use bronze for raw evidence, silver for validated and conformed Delta data, and gold for facts, dimensions, and aggregates.

Build metadata-driven pipelines using lookup, ForEach, parameters, dynamic expressions, conditions, retries, schedules, events, and failure paths.

Use Dataflow Gen2 for Power Query shaping and notebooks for reusable PySpark transformations, joins, denormalization, grouping, and aggregation.

Handle duplicates, missing data, invalid types, late records, schema drift, quarantine, reconciliation, and quality metrics deterministically.

Create and secure OneLake shortcuts; select mirroring when managed database/catalog synchronization better matches the source.

Streaming engineering with Eventstreams, Eventhouse, Spark, and KQL

Weeks 6-7

Reason from event time, latency, state, and operations. Compare native Eventhouse ingestion, standard and accelerated OneLake shortcuts, Eventstream transformations, and code-first Structured Streaming.

Define event ID, schema version, event versus ingestion time, allowed lateness, duplicate behavior, checkpoint, reject, and retention policy.

Use Eventstream to connect sources, manage fields, filter, enrich, aggregate in windows, derive streams, and route destinations.

Use Eventhouse and KQL for high-volume event, log, semi-structured, time-series, and operational analytics.

Write KQL that filters by datetime early, projects narrowly, summarizes with bin, and avoids unnecessary joins and dynamic expansion.

Build a Structured Streaming query with explicit schema, event-time watermark, state-aware transformation, Delta sink, and durable unique checkpoint.

Benchmark native tables, standard OneLake shortcuts, and query acceleration by latency, feature support, hot period, region, SSD, compute, and cost.

Security, governance, and lifecycle engineering

Weeks 8-9

Make releases reviewable and data paths defensible. Source control an item definition only when supported, promote through isolated stages, and treat data/checkpoint migration as separate from item rollback.

Use Fabric Git integration with branches and reviews for supported items, keeping credentials and environment-specific values out of source.

Implement database projects for supported Warehouse schema lifecycle and validate build, comparison, deployment, and rollback.

Promote supported items through development, test, and production deployment-pipeline stages with deliberate bindings and rules.

Apply workspace-, item-, row-, column-, object-, folder-, and file-level controls at the correct enforcement layer; add masking where appropriate.

Use sensitivity labels, endorsement, domains, OneLake catalog, lineage, owners, and audit evidence as complementary governance controls.

Run positive and negative access tests through every actual OneLake, shortcut, Spark, SQL, KQL, item, and downstream path.

Monitoring, optimization, projects, and exam readiness

Weeks 10-12

Finish with failure evidence and measured tuning. Successful execution is not enough: prove freshness, data quality, security, recovery, performance, capacity efficiency, release safety, and cleanup.

Use Monitoring hub, workspace monitoring, pipeline/Dataflow run details, notebook output, Spark UI, Eventstream insights/logs, Eventhouse monitoring, query insights, and Capacity Metrics.

Diagnose pipeline, Dataflow Gen2, notebook, Eventstream, Eventhouse, T-SQL, and shortcut failures from the narrowest trustworthy evidence.

Optimize Delta files, joins, shuffle, skew, Spark memory, pipeline movement, Warehouse SQL, KQL reduction, Eventhouse cache/retention, and query acceleration.

Alert on actual freshness, quality, backlog, watermark, destination, query, throttling, and capacity objectives—not only scheduler starts.

Complete both substantial DP-700 projects with security, deployment, quality, failure injection, performance, capacity/cost, runbooks, and cleanup evidence.

Take mixed original practice, explain every distractor, retrieve key distinctions with flashcards, revisit weak objectives, and recheck the official guide.

PrepKloud learning surfaces

Original practice

Work through 25 varied scenarios balanced across the three current 30-35% domains.

Configure DP-700 practice →

Retrieval flashcards

Recall service boundaries, security layers, loading patterns, stream semantics, monitoring evidence, and tuning decisions.

Open DP-700 flashcards →

Portfolio projects

Build a governed medallion lakehouse platform and a real-time event engineering and operations platform.

Explore DP-700 projects →

Long-form guide

Study the complete domain strategy, architecture choices, troubleshooting loop, and readiness plan.

Read the DP-700 guide →

Career research

Compare Fabric data engineer requirements with the evidence produced by the two projects.

Explore data engineering jobs →

Official Microsoft references

DP-700 study guide

The source of current objectives, exact domain ranges, audience profile, and change log.

Open the study guide →

Fabric Data Engineering

The official entry point for Lakehouse, Spark, notebooks, jobs, medallion design, and engineering operations.

Study Data Engineering →

OneLake

Review shared storage, shortcuts, mirroring choices, security, and cross-engine access.

Study OneLake →

Fabric Data Factory

Review pipelines, Dataflow Gen2, connectors, orchestration, triggers, monitoring, and troubleshooting.

Study Data Factory →

Real-Time Intelligence

Review Eventstreams, Eventhouse, KQL, OneLake shortcuts, query acceleration, dashboards, and alerts.

Study Real-Time Intelligence →

Fabric monitoring

Ground item logs, monitoring Eventhouse, KQL investigations, alerts, and operational visibility.

Study workspace monitoring →

Frequently asked questions

Is DP-700 active in August 2026?

Yes. DP-700 is active as of August 19, 2026. The current English skills measured took effect July 21, 2026. Localized exams can update later, so verify the study guide and exam page for your date and language.

How should study time be allocated?

Allocate approximately one third to each official domain, then use cross-domain cases. A technically correct transformation can still fail because security, deployment, monitoring, capacity, or replay behavior was omitted.

Must I know SQL, PySpark, and KQL?

Yes. The current audience profile names all three. Practice SQL for relational transformations and Warehouse work, PySpark for scalable lake engineering, and KQL for event and operational analytics.

Does DP-700 include real-time engineering?

Yes. The blueprint includes streaming loading patterns, engine selection, native Eventhouse tables and OneLake shortcuts, query acceleration, Eventstreams, Spark Structured Streaming, KQL, and window functions.

Does PrepKloud reproduce live exam questions?

No. These questions, flashcards, projects, roadmap, and guide are original educational content based on public objectives and official documentation. Avoid dumps and recalled questions; use hands-on engineering and the current Microsoft guide.

Turn the blueprint into operational evidence

Read the documented behavior, retrieve it from memory, make an architecture decision, build with synthetic data, inject one failure, and record security, deployment, quality, performance, capacity, cost, and recovery results.

Start practice Review flashcards Build projects Explore jobs