Microsoft Certified: Fabric Data Engineer Associate — Practice Set 1

Question 1

You are setting up a Microsoft Fabric workspace for a new data engineering project. You need to configure the Spark environment to use a specific runtime version and enable autoscaling. Where should you configure these settings?

Accepted Answer

Spark workspace settings in Microsoft Fabric allow you to configure the Spark runtime version, node families, autoscaling behavior, and other Spark-specific options for all notebooks and Spark jobs in the workspace.

Question 2

You are managing a Microsoft Fabric workspace. You need to ensure that all data stored in OneLake for this workspace is encrypted and access-controlled at the folder level. Which setting should you configure?

Accepted Answer

OneLake workspace settings allow you to configure OneLake-specific options including folder-level access controls, integration with external storage, and OneLake security settings for the workspace.

Question 3

Your organization requires that all Microsoft Fabric workspaces in a department share common data governance policies and sensitivity label defaults. What should you configure?

Accepted Answer

Fabric domains group related workspaces under shared governance policies, delegated administration, and sensitivity label defaults. Assigning workspaces to a domain via domain workspace settings enables consistent governance across all workspaces in the department.

Question 4

You are implementing security for a Microsoft Fabric data warehouse. You need to ensure that users in the Sales role can only see rows related to their own region, and cannot view salary columns. Which two security features should you implement?

Accepted Answer

Row-level security (RLS) restricts which rows a user can see based on filter predicates, while column-level security (CLS) restricts access to specific columns. Together they address both requirements: region-based row filtering and salary column hiding.

Question 5

You are a data engineer at a healthcare organization. Your Microsoft Fabric workspace contains patient data classified as highly sensitive. Regulatory requirements mandate that all access to this data be logged for audit purposes. What should you implement?

Accepted Answer

Microsoft Fabric audit logs record user activities including data access events across Fabric items. For regulatory compliance in healthcare, audit logs provide the required evidence of who accessed what data and when. Sensitivity labels classify data but do not log access events.

Question 6

You need to implement lifecycle management for a Microsoft Fabric solution. Which two features support version control and environment promotion for Fabric items? Choose 2.

Accepted Answer

Git integration enables source control for workspace items, allowing version tracking, branching, and rollback. Deployment pipelines support structured promotion of items through Development, Test, and Production stages. Together these two features implement a complete lifecycle management solution for Fabric.

Question 7

You need to orchestrate a data pipeline in Microsoft Fabric that runs a notebook, waits for it to complete, then triggers a Dataflow Gen2 refresh. Which orchestration approach should you use?

Accepted Answer

Microsoft Fabric pipelines support activity sequencing with dependency control. You can add a notebook activity followed by a Dataflow Gen2 activity with a success dependency, ensuring the dataflow only runs after the notebook completes successfully.

Question 8

You need to implement dynamic data masking in a Microsoft Fabric data warehouse so that call center agents can see only the last four digits of customer credit card numbers. How should you implement this?

Accepted Answer

Dynamic data masking in Microsoft Fabric warehouses allows you to define masking rules on columns so that non-privileged users see masked values (e.g., 'XXXX-XXXX-XXXX-1234') while privileged users see the full data. This is more flexible than column-level security, which would hide the column entirely.

Question 9

You are designing a data loading solution in Microsoft Fabric. You receive new transaction files every hour and need to process only the files that arrived since the last run. Which loading pattern should you implement?

Accepted Answer

An incremental load pattern using a watermark (last processed timestamp or file marker) processes only new data since the previous run, reducing redundant processing and resource consumption for frequently arriving data.

Question 10

You need to ingest streaming sensor data into a Microsoft Fabric Eventhouse for real-time analysis. Which service should you use to ingest and route the streaming events?

Accepted Answer

Microsoft Fabric Eventstream is the streaming ingestion service that connects to real-time event sources (such as Azure Event Hubs, IoT Hub, or Kafka) and routes streaming data to destinations including Eventhouse, lakehouses, and other Fabric items.

Question 11

You need to transform large volumes of batch data in Microsoft Fabric using distributed processing with Python. Which tool should you use?

Accepted Answer

Microsoft Fabric notebooks with PySpark leverage Apache Spark's distributed processing capabilities for large-scale batch data transformation using Python. PySpark is ideal for complex transformations on large datasets that exceed what single-node processing can handle.

Question 12

You are building a data engineering solution in Microsoft Fabric. You need to reference data stored in Amazon S3 from a lakehouse without copying it into OneLake. What should you create?

Accepted Answer

OneLake shortcuts support external data sources including Amazon S3. A shortcut creates a virtual path in the lakehouse that references the S3 data in place, allowing queries and Spark jobs to access it without physically moving the data into OneLake.

Question 13

You are a data engineer at a logistics company. Your streaming data pipeline receives GPS coordinates from delivery vehicles every 10 seconds. You need to calculate the average speed for each vehicle over the last 5 minutes of data. Which approach should you implement?

Accepted Answer

KQL in Microsoft Fabric Eventhouse supports time-series windowing functions including tumbling, sliding, and session windows. For real-time aggregation over a rolling 5-minute window per vehicle, KQL windowing queries running against the Eventhouse provide low-latency, stream-native computation.

Question 14

You need to choose the right tools to transform batch data in Microsoft Fabric. Your transformation requires complex SQL joins across multiple lakehouse tables and also Python-based data quality checks. Which two tools are most appropriate? Choose 2.

Accepted Answer

T-SQL is ideal for complex join-based transformations in a Fabric warehouse or via the lakehouse SQL analytics endpoint. PySpark notebooks are best suited for Python-based data quality logic and large-scale distributed transformations. Both can be orchestrated together in a Fabric pipeline.

Question 15

You need to replicate data from an Azure SQL Database into a Microsoft Fabric lakehouse in near-real-time without building custom ETL pipelines. Which Microsoft Fabric feature should you use?

Accepted Answer

Microsoft Fabric mirroring continuously replicates data from supported sources including Azure SQL Database, Azure Cosmos DB, and Snowflake into OneLake as Delta Parquet tables in near-real-time, without requiring custom ETL pipelines.

Question 16

You are processing streaming data from multiple IoT sensors in Microsoft Fabric. Sensor readings sometimes arrive out of order due to network latency, up to 30 seconds late. You need to ensure late-arriving data is included in the correct time window aggregations. What should you configure?

Accepted Answer

Spark Structured Streaming supports configurable watermarks that allow a defined late-arrival tolerance window. Setting a 30-second watermark means the streaming engine will wait up to 30 seconds for late data before finalizing a window aggregation, ensuring late-arriving sensor readings are included in the correct windows.

Question 17

You are designing a dimensional data model in Microsoft Fabric. Before loading data into the fact table, you need to prepare dimension tables with surrogate keys and standardized attribute formats. What loading pattern step does this represent?

Accepted Answer

Preparing data for loading into a dimensional model involves standardizing dimension attributes, generating surrogate keys, handling slowly changing dimensions, and ensuring referential integrity before the fact table load. This is a distinct and critical step in the dimensional loading pattern.

Question 18

You are a data engineer at a retail company. A nightly data pipeline in Microsoft Fabric began failing intermittently last week. The pipeline includes a copy activity followed by a notebook. You need to identify the root cause. What should you do first?

Accepted Answer

The pipeline run history in the Microsoft Fabric portal shows the execution status of each activity, including detailed error messages, timestamps, and duration. Reviewing this is the first step to identifying whether the copy activity, the notebook, or an upstream dependency is the root cause of the failure.

Question 19

You need to set up proactive notifications in Microsoft Fabric so that your team is alerted when a data pipeline fails. What should you configure?

Accepted Answer

Microsoft Fabric supports configurable alerts on monitored items including pipelines. You can set alerts to trigger notifications via email or Teams when a pipeline run fails, succeeds, or meets other conditions.

Question 20

You are optimizing a Microsoft Fabric lakehouse table that is used in frequent analytical queries. After multiple data loads, query performance has degraded. What should you do to optimize the table?

Accepted Answer

Over time, frequent writes to a Delta table create many small Parquet files that degrade query performance. Running OPTIMIZE compacts these into larger files, and VACUUM removes stale Delta version files, restoring query performance for the lakehouse table.

Question 21

You need to monitor semantic model refresh operations in Microsoft Fabric to ensure they complete within the defined SLA window. Which monitoring approach should you use?

Accepted Answer

The Microsoft Fabric monitoring hub provides a centralized view of all activity across Fabric items including semantic model refreshes, showing start time, duration, status, and error details. This is the recommended tool for tracking refresh SLA compliance.

Question 22

You are troubleshooting a Dataflow Gen2 that is failing intermittently in Microsoft Fabric. Which two actions should you take to identify the root cause? Choose 2.

Accepted Answer

Reviewing the Dataflow Gen2 refresh history reveals the failure pattern and error messages. Enabling diagnostic logging provides detailed traces including query evaluation steps, connector errors, and timeout information. Together these two actions give the most complete picture of the root cause.

Question 23

Your Spark notebook job in Microsoft Fabric is running slowly. The Spark UI shows that one executor is processing 90% of the data while others are idle. What is the most likely cause and resolution?

Accepted Answer

When one Spark executor processes significantly more data than others, the cause is typically data skew — certain partition key values have disproportionately large data volumes. The resolution is to repartition by a more evenly distributed key or use salting techniques to spread the skewed data across multiple partitions.

Question 24

You are a data engineer at a financial services firm. Your Eventhouse KQL queries are timing out during peak trading hours. You suspect that Eventstream is delivering data faster than the Eventhouse can ingest it. What should you investigate and optimize?

Accepted Answer

Eventhouse performance during high-throughput periods can be improved by tuning ingestion batching policies (to reduce ingestion latency), increasing partition counts for high-volume tables, and expanding the hot cache window so that frequently queried recent data is served from memory rather than cold storage.

Question 25

You need to monitor data ingestion pipelines in Microsoft Fabric. Which two metrics should you track to ensure pipelines are performing within acceptable thresholds? Choose 2.

Accepted Answer

Monitoring pipeline run duration and row counts helps identify performance degradation and unexpected data volume changes. Tracking success and failure rates over time reveals reliability trends and helps prioritize incident response. Both are core operational metrics for data pipeline monitoring.

DP-700 — Fabric Data Engineer Associate