DP-700 Domains Explained With Weightings

The short version

DP-700 has three domains, and Microsoft gives each a weighting of 30-35%. That makes the exam balanced: you cannot pass on pipelines alone. The lists below follow the study guide labeled “Skills measured as of October 19, 2026”. The traps and practice tips are our own advice, not Microsoft’s wording.

Domain 1: Implement and manage an analytics solution (30-35%)

This domain is about the environment around your data work: settings, change control, security, and orchestration.

Configure workspace settings

Microsoft lists Spark, domain, OneLake, and Apache Airflow workspace settings. In practice you decide the default Spark pool and runtime, how a workspace belongs to a domain, and what OneLake options apply.

Trap: Fabric offers prewarmed starter pools, but they only apply to certain configurations. If a scenario needs a custom setup, ask what that does to session start time.

Practice: Open the workspace settings and read every Spark option once.

Lifecycle management

Version control (Git integration), database projects, and deployment pipelines. Expect questions about what you can commit, what an update does, and how content moves between stages.

Trap: Mixing up Git integration and deployment pipelines. Git tracks item definitions in a repository. Deployment pipelines promote content through stages. A scenario about “moving to Test” is usually the second.

Practice: Connect a workspace to a repository, commit, change something in the repository, and update. Then build a three-stage deployment pipeline and note what it asks you to configure per stage.

Security and governance

This is the longest list in the domain: workspace-level and item-level access, row-level, column-level, object-level, and folder or file-level access, dynamic data masking, sensitivity labels, endorsement, audit logs, and OneLake security.

Trap: Choosing the broadest tool. If the requirement says “only the finance rows”, a workspace role is too wide and row-level security fits. Also remember that masking hides values from readers; it does not stop a user from having access.

Practice: For each tool, write one sentence starting “Use this when…”. Test row-level security and masking on a small warehouse table.

Orchestrate processes

Choose between Dataflow Gen2, a pipeline, and a notebook, design schedules and event-based triggers, and use parameters and dynamic expressions in notebooks and pipelines.

Trap: Treating all three tools as interchangeable. Low-code transformation points to Dataflow Gen2, control flow across steps points to a pipeline, and custom code at scale points to a notebook.

Practice: Build one pipeline that calls a notebook with a parameter, then triggers on a schedule.

Domain 2: Ingest and transform data (30-35%)

Design and implement loading patterns

Full and incremental loads, preparing data for a dimensional model, and a loading pattern for streaming data.

Trap: Picking incremental loading when the source has no reliable way to find changes. Ask what tells you a row is new or changed: a watermark column, change tracking, or something else.

Practice: Implement a watermark-based incremental load, then a dimension load with a surrogate key.

Ingest and transform batch data

Choose a data store, choose between Dataflows Gen2, notebooks, KQL, and T-SQL, create shortcuts, implement mirroring, ingest with pipelines, transform with PySpark, SQL, and KQL, denormalize, group and aggregate, and handle duplicate, missing, and late-arriving data.

Trap: Shortcut and mirroring confusion. A shortcut points to data where it already lives, without copying it. Mirroring keeps a replicated copy of an external database in OneLake. Also note that some shortcut types are read-only, such as Amazon S3.

Practice: Create a shortcut to another lakehouse, and write the same aggregation in PySpark and SQL.

Ingest and transform streaming data

Choose a streaming engine, choose between native tables and shortcuts in Real-Time Intelligence, choose between Query acceleration for OneLake shortcuts and standard shortcuts, process data with Eventstream, Spark structured streaming, and KQL, and create windowing functions.

Trap: Window types. Tumbling windows do not overlap; each event belongs to exactly one. Hopping windows overlap when the hop is shorter than the window. Sliding and session windows have their own rules.

Practice: Write one KQL query and one Spark streaming query that each use a window, and say aloud why you picked the type.

Domain 3: Monitor and optimize an analytics solution (30-35%)

Monitor Fabric items

Monitor data ingestion, transformation, and semantic model refresh, and configure alerts.

Trap: Looking in the wrong tool. Learn which view shows run history, which shows Spark detail, and which shows capacity use.

Practice: Run a pipeline and a notebook, then find each in the monitoring hub and read the details.

Identify and resolve errors

Pipeline, Dataflow Gen2, notebook, Eventhouse, Eventstream, T-SQL, and OneLake shortcut errors. Questions give a symptom and ask for the likely cause or the quickest fix.

Trap: Fixing the symptom you can see instead of reading the exact message. The deciding clue is often a single word, such as “permission” or “schema”.

Practice: Break something on purpose (a wrong parameter, a missing permission) and trace the error.

Optimize performance

Lakehouse tables, pipelines, data warehouses, Eventstream and Eventhouse, Spark, and queries. Think table maintenance, partitioning, file size, and query shape. The lists are broad, so learn the idea behind each lever before you memorize names.

Trap: Optimizing before measuring. Scenarios usually say what is slow; match the lever to that.

Practice: Compare the run time of a query before and after maintenance on a table with many small files.

More guides for DP-700

Next step

Take the sample questions to see these domains in scenario form. When you are ready to check yourself across all three, our DP-700 practice tests report your score per domain after each attempt.

Try the DP-700 practice exams