Microsoft Fabric Data Engineer Interview questions

1. How would you handle multi-domain data governance in Fabric?

To handle multi-domain data governance, Use domains in OneLake, apply Purview for classification and enforce access policies per workspace/domain.

2. State the main components of Microsoft Fabric

Data Factory
Synapse Data Engineering
Synapse Data Science
Synapse Data Warehouse
Real-Time Analytics
Power BI
Data Activator
OneLake (Unified storage)

3. Explain about OneLake

OneLake is a single, unified logical data lake that stores all data in Microsoft Fabric in Delta Lake format.

4. How does Fabric support lake house architecture?

Microsoft Fabric supports lake house architecture by combining the scalability of data lakes and the structure of data warehouses using Delta Lake and Apache Spark.

5. What languages does Fabric support?

T-SQL
PySpark
SQL
KQL (Kusto Query Language)
DAX
Power Query (M)

6. What is Lakehouse and Warehouse in Fabric?

Lakehouse combines the capabilities of data warehouses and data lakes to store structured and unstructured data in a Delta Lake format.

Warehouse is a fully managed, scalable, and performant relational database designed for analytical workloads.

7. How would you enable real-time alerting on stock price changes in Fabric?

Ingest data using Eventstream
Query using KQL
Create a Data Activator rule

8. What is a Notebook in Fabric?

Notebook is a collaborative document to run code like SQL, PySpark, etc. for data exploration, transformation, or machine learning.

9. What is the role of a Dataflow in Fabric?

The role of a Dataflow is used to ingest, transform, and load data into Fabric Lakehouse or Warehouse using Power Query.

10. Write the difference between Lakehouse and Warehouse in Fabric?

Lakehouse uses files (Delta format) which is better for big data whereas Warehouse uses tabular structures (T-SQL), suitable for BI-style reporting.

11. How do you ingest data in Microsoft Fabric?

Pipelines (Data Factory)
Dataflows Gen2
Notebooks

12. What is Delta Lake format and How do you handle schema evolution in Delta Lake?

Delta lake format is an open-source storage format for big data workloads that supports ACID transactions, schema enforcement, and time travel. Schema evolution can be handled by using “mergeSchema” or “autoMerge” options while writing data.

13. How do you troubleshoot a failed Spark job in Fabric?

Checking job logs and spark history server
Error messages in the Monitoring pane

14. What are Pipelines in Fabric?

Pipelines are orchestrated workflows to move and transform data, similar to Azure Data Factory pipelines.

15. How do you schedule and monitor data pipelines in Fabric?

Schedule and monitor data pipeline using triggers, activity monitoring, and alerts in the Data Factory component.

16. What is a Spark Job Definition in Fabric?

Spark job is a complete computation process that defines a job that runs PySpark code on a Spark compute environment.

17. What is a Lakehouse shortcut?

Lakehouse shortcut links to external data locations within OneLake, enabling shared access to data across domains.

18. How is data security handled in Fabric?

Microsoft Purview
Access policies
Role-based security

19. Explain medallion architecture in Fabric.

Medallion architecture uses three structured layers: Bronze (raw data), Silver (cleansed), and Gold (aggregated).

Each layer serves a purpose in data refinement and performance optimization within Lakehouse environments.

This layered approach enhances scalability, data quality, and business analytics.

20. What is time travel in Delta Lake?

Time travel enables access to historical versions of data using version numbers or timestamps.

This is useful for auditing, debugging, or recovering previous datasets.

Fabric supports Delta Lake time travel through built-in version control features.

21. What are Semantic Models in Fabric?

Semantic Models are tabular data models used for reporting and analytics, especially in Power BI.

They define relationships, measures, and calculated columns for business analysis.

These models offer a consistent, reusable layer for data interpretation across reports.

22. What is Direct Lake Mode in Power BI?

Direct Lake Mode allows Power BI to directly query Delta Lake data in OneLake without importing it.

It combines the performance of Import with real-time freshness of DirectQuery.

This mode ensures faster reports and up-to-date insights for large datasets.

23. State the difference between Dataflow Gen1 and Gen2.

Dataflow Gen1 was designed for Power BI, using Power Query for basic ETL operations.

Dataflow Gen2 is part of Microsoft Fabric, built on a Spark engine for advanced, scalable processing.

Gen2 supports enhanced performance, parallelism, and better integration with Lakehouse.

24. How do you implement CI/CD in Microsoft Fabric?

CI/CD in Fabric is enabled through Git integration with workspaces and deployment pipelines.

You can version-control notebooks, dataflows, semantic models, and pipelines.

This streamlines collaboration, testing, and automated deployments in the development lifecycle.

25. How can you handle slowly changing dimensions in Fabric?

Use Delta Lake’s MERGE operation to update and track changes in dimension records.

Versioning can also be applied to maintain historical changes over time.

This ensures consistency in reporting and temporal accuracy of dimension data.

26. How does Power BI integrate with Fabric?

Power BI connects natively to Fabric’s Lakehouse, Data Warehouse, and Dataflows.

It can access data via Direct Lake, DirectQuery, or Import modes.

This integration ensures a seamless data analytics experience from raw data to dashboards.

27. What is a Semantic Model in Power BI?

A Semantic Model defines tables, relationships, and business metrics for reporting.

It acts as an abstraction layer between raw data and Power BI visuals.

Semantic models help in building reusable and scalable data models for analytics.

28. Can Power BI connect to a Lakehouse?

Yes, Power BI supports direct connection to Lakehouse using Import, DirectQuery, or Direct Lake modes.

This enables high-performance access to structured data stored in OneLake.

It helps create real-time, data-rich reports from Lakehouse sources.

29. What is the difference between Direct Lake and Direct Query?

Direct Lake reads Delta tables directly in OneLake with low latency and no transformation layer.

Direct Query translates Power BI queries into SQL and queries the source at runtime.

Direct Lake offers better performance for large datasets with frequent report access.

30. How do you handle large datasets in Power BI with Fabric?

Use aggregations, composite models, and Direct Lake access to optimize performance.

Divide data into logical partitions and enable incremental refresh where applicable.

This allows handling of billions of rows with interactive report performance.

31. What are Metrics in Power BI Fabric?

Metrics are KPIs tracked using scorecards to monitor business performance against targets.

They are interactive, can be connected to semantic models, and auto-refresh with data updates.

Fabric enables centralized metric tracking across workspaces for visibility and accountability.

32. What is OneLake in Microsoft Fabric?

OneLake is a unified data lake that stores all data in Microsoft Fabric.

It uses Delta format and allows shared access across Lakehouse, Warehouse, and Power BI.

OneLake eliminates data silos and enables a single-source-of-truth architecture.

33. What kind of visualizations are supported in Power BI?

Charts
Graphs
KPIs
Maps
Custom visuals

34. What is Row-Level Security (RLS)?

Row-Level Security (RLS) restricts data access at the row level based on user roles.

It is implemented in the semantic model using DAX filters or role definitions.

This ensures users only see data relevant to their identity or permission level.

35. How do you publish Power BI reports in Fabric?

Create or upload reports in Power BI Desktop and publish them to Fabric workspaces.

Use the Power BI Service to manage, refresh, and share reports with authorized users.

Workspaces act as containers to organize and collaborate on reporting assets.

36. What is KQL and where is it used?

KQL (Kusto Query Language) is a read-only query language for querying log and telemetry data.

It is primarily used in Real-Time Analytics within Fabric, often with KQL databases.

KQL enables fast, scalable querying of structured, semi-structured, and streaming data.

37. How do you use Eventstreams in Fabric?

Eventstreams in Fabric collect, process, and distribute real-time event data from various sources.

It integrates with IoT devices, Event Hubs, and Azure services for low-latency processing.

Supports data transformations and routing to Lakehouse, KQL DBs, or Power BI dashboards.

38. How does Real-Time Analytics in Fabric work?

Real-Time Analytics uses KQL databases to ingest and store time-series or event data.

It supports stream ingestion, fast queries, and interactive visualizations in Power BI.

This enables monitoring and insights from live operational data with minimal latency.

39. What is Data Activator’s purpose?

Data Activator in Fabric detects patterns, thresholds, or anomalies in real-time data streams.

It triggers automated workflows or alerts when defined conditions are met.

Helps with real-time decision-making and response to business events.

40. How is monitoring done in Fabric?

Fabric provides built-in dashboards for monitoring Spark jobs, pipelines, and data activities.

Metrics like success/failure rates, runtimes, and throughput are tracked.

You can also export logs or integrate with Azure Monitor for centralized monitoring.

41. How would you build a data pipeline from an on-premise SQL Server to Fabric Lakehouse?

Use Data Gateway to connect the on-premise SQL Server with Microsoft Fabric.

Use Dataflow Gen2 or pipeline with Copy activity to move data to a Delta Lake format.

Data is then stored in a Lakehouse and can be processed further using notebooks or semantic models.

42. How would you migrate from Azure Synapse to Fabric?

Rebuild Synapse pipelines in Data Factory (Fabric) using equivalent copy and transformation activities.

Replace Synapse SQL pools with Lakehouses for storage and compute operations.

Move existing semantic models to Power BI to maintain reporting capabilities.

43. What are best practices for Lakehouse folder structure?

Organize data into Bronze, Silver, and Gold folders for medallion architecture.
Partition data by date or region to improve query performance.
Use consistent and meaningful naming conventions for tables and folders.

44. What’s the difference between batch and stream processing?

Batch: Processes data in bulk at scheduled intervals (e.g., hourly, daily).
Stream: Processes data continuously in near real-time as it arrives.
Use batch for historical data and stream for immediate insights or alerts.

45. How does Fabric handle data latency in real-time workloads?

Fabric minimizes latency using eventstream pipelines and low-latency KQL databases.

It processes streaming data with minimal delay, ensuring near real-time insights.

Optimizations in buffering and parallel processing enhance performance.

46. How is data lineage tracked in Microsoft Fabric?

Microsoft Purview is used to track end-to-end data lineage in Fabric.

It captures transformations, movements, and dependencies across data systems.

This enhances governance, auditing, and impact analysis for compliance.

47. What is Microsoft Purview used for in Fabric?

Data cataloging: Discover and organize data assets.
Classification: Identify sensitive data using built-in classifiers.
Governance: Define policies, ownership, and business terms.
Compliance: Monitor data usage for regulatory compliance.

48. How do you manage workspace roles in Fabric?

Fabric supports role-based access control (RBAC) to manage user permissions.

Admin: Full control over all assets and permissions.
Member: Can edit and share content.
Contributor: Can create content but not manage access.
Viewer: Read-only access to workspace items.

49. How do you implement data masking or anonymization?

Use transformations in Dataflows or pipelines to mask personally identifiable information (PII).

Apply DAX expressions in semantic models to obfuscate sensitive data in reports.

Masking ensures compliance with privacy and security policies.

50. How is data encryption handled in Fabric?

Data in Fabric is encrypted at rest using Microsoft-managed encryption keys.

Data in transit is protected using TLS/SSL protocols to ensure secure communication.

Compliance with enterprise-grade encryption standards is maintained across the platform.

Request more information

Microsoft Fabric Data Engineer Interview questions