What should you prepare for Azure Data Factory interviews?
Which ADF basics should I be ready to explain?
“What is Azure Data Factory?” was the most asked question, appearing in 4 of 42 interviews. Start there, then check whether you can explain an integration runtime and its role in connecting to data sources.
- Can you explain what ADF and Databricks each do, and when you would use each?
- Can you name commonly used activities and describe a pipeline using activities, parameters, variables, and triggers?
- Can you explain how you would transfer an ADF workflow to Airflow, identify gaps in your knowledge, and get started safely?
How much architecture should I cover?
Architecture and migration appeared in 13 of 42 interviews, more than any other topic. These questions covered whole data solutions, not just individual ADF features.
Can you describe a retail pipeline that takes data from multiple sources, uses ADF and Databricks, and loads results into Delta Lake? Explain how you would handle performance and failures, not just which tools you would use.
Also check whether you can compare platforms and plan a cross-cloud migration. Can you explain the security, governance, and ownership choices in your design?
Can I explain safe loading and recovery?
Data ingestion appeared in 8 of 42 interviews. Check both how data arrives and how you would confirm that the result is complete. Include what happens when a run needs to be repeated.
- How would you select source files and use metadata to control ingestion?
- How would you use change data capture (CDC) or watermarks for incremental loading and handle a rerun safely?
- If ADF runs succeed but report records are missing, how would you investigate and decide whether publication is safe?
- How would you handle incomplete batches, set up failure alerts, and document recovery steps?
Should I review Fabric and Power BI too?
Fabric and Power BI appeared in 8 of 42 interviews. Can you choose a Power BI storage mode for Fabric OneLake data and explain your choice?
Check whether you can describe the roles of Fabric pipelines, notebooks, and warehouses. Can you also explain SQL endpoints, semantic models, and how you would plan capacity?
What code should I be able to write?
Writing code was part of 3 of 42 interviews. The listed tasks used PySpark. Check whether you can write each workflow from reading the data to saving the result.
- Read a large CSV file from Azure Data Lake Storage.
- Remove rows with missing values in a chosen column, calculate a grouped average, and write a partitioned Delta table with schema enforcement.
- Inner join customer and transaction Parquet datasets on
customer_id, calculate total transaction amounts per customer, and save CSV output.
Who gets asked Azure Data Factory
The 20 most asked Azure Data Factory questions
Counted by interview, not by mention: a question asked twice in one conversation counts once.
What is Azure Data Factory?
Asked in 4 of 42 interviews · last asked Aug 25 · Concept · Junior
Show answer
Azure Data Factory (ADF) is a managed cloud service for moving and integrating data. It uses pipelines to copy data between systems and coordinate processing tasks. It supports ETL and ELT by running data flows or calling tools such as Azure Databricks. You can schedule pipelines and monitor their runs.
- Pipelines group activities such as copying data, transforming data, or running notebooks.
- Connectors support cloud and on-premises data sources.
- Triggers start pipelines on schedules or in response to events.
- ADF coordinates data processing; it is not a database or data warehouse.
What the interviewer checks: The interviewer checks whether you understand ADF’s role in moving, transforming, and coordinating data.
Other ways it was asked
- “How does Azure Data Factory (ADF) work?”
- “Please explain Azure Data Factory.”
- “What is Azure Data Factory used for in a data engineering workflow?”
What is an integration runtime in Azure Data Factory?
Asked in 3 of 42 interviews · last asked Oct 7 · Concept · Junior
Show answer
An integration runtime (IR) is the compute infrastructure Azure Data Factory uses to move data, run data flows, and dispatch activities. Azure IR is managed by Azure. Self-hosted IR runs on your Windows machines to reach on-premises or private-network data. Azure-SSIS IR runs SQL Server Integration Services (SSIS) packages.
- Choose the IR type based on network access and the work it must run.
- Azure IR can also access supported private resources through a managed virtual network and managed private endpoints.
- Create an IR under Manage > Integration runtimes, then select its type and configure it.
- For self-hosted IR, install the runtime on a Windows machine and register it using an authentication key.
What the interviewer checks: The interviewer checks whether you understand the three IR types and can choose one for the workload.
Other ways it was asked
- “How do you set up an integration runtime in Azure Data Factory?”
- “What is a self-hosted integration runtime in Azure Data Factory, and when would you use it?”
- “What are the different types of Azure Data Factory integration runtime?”
What roles do Azure Data Factory and Databricks play, and when would you use each?
Asked in 2 of 42 interviews · last asked Aug 14 · Scenario · Mid-level
Show answer
Azure Data Factory mainly moves data and coordinates pipeline steps across systems. Azure Databricks mainly processes and analyzes data using scalable compute, including Apache Spark. Use Data Factory for connector-based data movement and scheduling, and Databricks for complex transformations or machine learning. They can work together: Data Factory can run Databricks jobs.
- Data Factory offers many connectors and low-code pipelines; it also supports transformations through mapping data flows.
- Databricks supports code-heavy processing, streaming, and machine learning.
- Databricks also supports job scheduling, so Data Factory is not always needed.
- Use both when one pipeline must move data and coordinate Databricks processing with other services.
What the interviewer checks: The interviewer checks whether you can separate data movement and coordination from processing, while recognizing overlapping features.
Other ways it was asked
- “What are Databricks and Azure Data Factory, and what are their primary roles?”
- “Why might you use Azure Data Factory instead of implementing data processing in Databricks or data lake jobs?”
How would you transfer orchestration experience from Azure Data Factory to Airflow, identify learning gaps, and become productive safely?
Asked in 2 of 42 interviews · last asked Sep 22 · Scenario · Mid-level
Show answer
I would transfer my ADF skills in task dependencies, retries, monitoring, and safe reruns to Airflow. I would identify gaps by building a small Python DAG and testing scheduling, failures, and backfills. Spark needs separate learning because it processes data, while Airflow coordinates tasks. I would start in a sandbox, then use reviewed changes and a small production rollout.
- Map orchestration concepts, not features one-to-one; Airflow uses Python-defined DAGs rather than ADF’s visual pipelines.
- Learn Airflow data intervals, catchup, backfills, connections, and worker execution.
- For Spark, learn lazy execution, partitions, and shuffles; orchestration experience alone does not cover distributed processing.
- Make reruns safe, validate outputs, limit concurrency, and prepare monitoring and rollback before production.
What the interviewer checks: The interviewer checks whether you can reuse existing skills, recognize learning gaps, and learn without risking delivery.
Other ways it was asked
- “Which data orchestration skills transfer from Azure Data Factory to Airflow and Spark, where would you expect learning gaps, and how would you become productive without creating delivery risk?”
What are some commonly used activities in Azure Data Factory?
Asked in 2 of 42 interviews · last asked Oct 7 · Tooling · Junior
Show answer
Five commonly used Azure Data Factory activities are Copy Data, Data Flow, Lookup, ForEach, and Execute Pipeline. Copy Data moves data, and Data Flow transforms it. Lookup reads values for later steps, ForEach repeats activities for each item, and Execute Pipeline runs another pipeline.
- Copy Data moves data between supported sources and destinations.
- Data Flow supports visual data transformations using managed Spark compute.
- Lookup returns a first row or multiple rows, often used as settings or a work list.
- ForEach supports sequential or parallel execution; Execute Pipeline helps reuse pipeline logic.
What the interviewer checks: The interviewer checks whether you can choose suitable activities for moving, transforming, and coordinating data.
Other ways it was asked
- “Which Azure Data Factory activities would you use to build a data pipeline?”
- “What are five commonly used activities in Azure Data Factory?”
Write a PySpark script to read a large CSV file from Azure Data Lake Storage.
Asked in 2 of 42 interviews · last asked Aug 22 · Write code · Mid-level
Show answer
I use Spark’s CSV reader with an explicit schema and an abfss path for Azure Data Lake Storage Gen2. Spark processes the data across workers without loading everything into the driver. The Spark environment must have the Azure storage connector, authentication, and read permissions configured.
- Use a schema that matches the CSV columns and their order.
- An explicit schema avoids the extra scan needed to infer column types.
- Avoid collect() and toPandas() for large datasets.
- FAILFAST stops processing when Spark encounters a malformed record.
Pythonfrom pyspark.sql import SparkSession spark = SparkSession.builder.appName("ReadADLSCSV").getOrCreate() # Storage authentication must already be configured. path = "abfss://raw@myaccount.dfs.core.windows.net/sales/large.csv" schema = "sale_id BIGINT, sale_date DATE, sales_amount DOUBLE" df = (spark.read.schema(schema) .option("header", True) .option("mode", "FAILFAST") .csv(path)) df.show(5)What the interviewer checks: The interviewer checks whether you can access ADLS, set a schema, and read large files without overloading the driver.
Other ways it was asked
- “Write a PySpark script to read a large batch of JSON files from an Azure Data Lake Gen2 container, validate the JSON schema, write to a Delta Lake table partitioned by a column, and include error handling for schema mismatches with a summary of invalid records.”
- “Write a PySpark script to read a large batch of JSON files from an Azure Data Lake Gen2 container, validate the JSON schema, write to a Delta Lake table partitioned by a column, and handle schema mismatches by summarizing invalid records.”
- “Write a PySpark script to read a parquet file from Azure Data Lake, filter rows where the sales amount is greater than 5,000, and save the result as a new CSV file back to Azure Data Lake. Ensure your solution is efficient and handles large datasets.”
Which Power BI storage mode would you use for Fabric OneLake data, and why?
Asked in 2 of 42 interviews · last asked Sep 16 · Concept · Mid-level
Show answer
I would use Direct Lake for supported Delta tables in Fabric OneLake. It reads data directly from OneLake without creating a separate Import-mode copy. It loads needed columns into memory for fast queries and avoids full import refreshes.
- Direct Lake requires a Fabric capacity and supported Delta tables.
- Unlike DirectQuery, it normally reads OneLake files rather than querying a SQL endpoint.
- It still uses memory; no import does not mean no caching.
What the interviewer checks: The interviewer is checking whether you can distinguish Direct Lake from Import and DirectQuery.
Other ways it was asked
- “Which connection mode would you use for a Power BI report whose data is stored in Microsoft Fabric, and why?”
- “Which Power BI storage mode lets you work with Fabric OneLake data without importing it into the semantic model?”
How would you investigate missing report records after successful ADF runs and decide whether publication is safe?
Asked in 2 of 42 interviews · last asked Sep 29 · Scenario · Senior
Show answer
A successful ADF run does not prove the data is complete. I would compare the same batch from the source through staging, final tables, and Power BI to find where records disappear. I would publish only after completeness checks pass; unexplained missing transactions or failed business checks would block publication.
- Compare row counts, transaction keys, and business totals for the same date range and cutoff.
- Check ADF copy metrics, skipped rows, watermarks, time zones, filters, and joins.
- Check Power BI refresh coverage, incremental refresh partitions, relationships, report filters, and row-level security.
- Use agreed validation rules; keep the last validated report available and clearly label its data date.
What the interviewer checks: The interviewer checks whether you can trace missing data across systems and separate technical success from business validity.
Other ways it was asked
- “An ADF pipeline reports success, but the Power BI report contains fewer records than the source extract. What would you check first, and how would you decide when the report is safe to use?”
- “If the ADF pipeline and Power BI refresh both succeed but a business user says yesterday’s customer transactions are incomplete, what checks would you run before calling the batch valid, and which failure would stop publication?”
How would you design a data ingest and transformation pipeline for a retail company using Azure Data Factory and Azure Databricks to process data from multiple sources and load the results into a Delta Lake, maximizing performance and handling failures effectively?
Asked in 2 of 42 interviews · last asked Aug 11 · Design · Senior
Show answer
I would use ADF to ingest incremental data from each source into raw storage in ADLS. Databricks would validate, clean, and enrich it, then write curated Delta tables. I would parallelize independent loads and use Delta MERGE with stable keys to make reruns safe. ADF would retry temporary failures, quarantine invalid records, and alert on failed runs.
- Use CDC or watermarks; advance ingestion watermarks only after raw data is safely stored.
- Filter early, broadcast small lookup tables, and compact small Delta files; partition only when queries benefit.
- Use autoscaling job compute and terminate it after processing; limit concurrency to protect source systems.
- Use self-hosted integration runtime for on-premises sources, private networking, managed identities where supported, and Key Vault.
What the interviewer checks: The interviewer checks whether you can connect ingestion, Spark performance, cost control, security, and safe recovery.
Other ways it was asked
- “You're tasked with designing an ETL pipeline using Azure Databricks to process incremental data loads from Azure Data Lake. The data contains multiple large Parquet files, and you need to perform transformations such as filtering, aggregations, and data enrichment before writing the output back to Azure SQL database. How would you ensure efficient data processing, optimize costs, and handle potential failures in the pipeline?”
- “How would you structure an Azure Data Factory pipeline to optimize performance and ensure fault tolerance when ingesting large volumes of data from multiple sources, transforming it, and loading it into Azure SQL Database?”
- “How would you design an Azure Data Factory pipeline for data ingestion from on-premise SQL Server, processing in Azure Databricks, and loading into Azure Synapse Analytics to ensure security, scalability, and performance?”
Why do you need Azure when you are already using Databricks to process and store data?
Asked once · last asked Aug 26 · Concept · Mid-level
Show answer
Azure is the cloud platform; Azure Databricks is a service within it for processing and analyzing data. Azure Data Lake Storage stores the files behind typical Databricks tables, while Databricks provides compute and table management. Azure Data Factory can move data and run workflows, but it is optional. You need Databricks only if its processing features fit your workload.
- Databricks tables do not replace the underlying cloud storage.
- Azure Data Lake Storage stores data; it does not run Spark transformations.
- Azure Databricks depends on Azure, but Databricks also runs on other supported clouds.
- Data Factory is optional because Databricks can also schedule and run workflows.
What the interviewer checks: The interviewer checks whether you can separate cloud infrastructure, storage, processing, and workflow management.
Other ways it was asked
- “Why do you need Azure storage if you can just store data in Databricks tables?”
- “When you are already using Azure for the data lake and the middle layer architecture, why do you specifically need Databricks?”
- “Why do you need Databricks if you are already using Azure Data Lake Storage for your middle layer?”
Which Microsoft Azure certifications are relevant to designing and implementing enterprise-scale Azure data architectures, such as Azure Solutions Architect Expert, Azure Data Engineer Associate, Azure Administrator Associate, or Azure AI Engineer Associate?
Asked once · last asked Jun 10 · Concept · Mid-level
Show answer
Azure Solutions Architect Expert is the strongest match for enterprise-wide Azure architecture. Azure Data Engineer Associate covered data pipelines and storage, but retired in March 2025. Azure Administrator Associate supports the infrastructure skills needed to run these systems. Azure AI Engineer Associate focused on AI solutions, not general data architecture.
- Solutions Architect Expert requires Azure Administrator Associate and passing AZ-305.
- Administrator skills include identity, networking, storage, compute, and governance.
- Fabric Data Engineer Associate is a data engineering option, but focuses on Microsoft Fabric.
- Choose architecture certifications for broad design roles and specialist certifications for specific workloads.
What the interviewer checks: The interviewer checks whether you can match certifications to architecture, data engineering, infrastructure, and AI responsibilities.
What types of resources would you provision in Azure for a data platform?
Asked once · last asked Sep 16 · Scenario · Mid-level
Show answer
I would provision storage, data processing, orchestration, and a database or warehouse for reporting. Common choices are Azure Data Lake Storage, Azure Databricks, Azure Data Factory, and Azure SQL Database or Azure Synapse Analytics. I would also add security, networking, and monitoring resources. The exact services depend on data volume, latency, and cost.
- Use Data Lake Storage for files and a database or warehouse for reporting.
- Use Data Factory to move data and schedule workflows; choose compute for the required transformations.
- Use managed identities, Key Vault, and private endpoints to secure access.
- Use Azure Monitor for alerts; add Event Hubs if streaming ingestion is needed.
What the interviewer checks: The interviewer is checking whether you can choose services for a complete platform, including operations and security.
What do you understand by a tumbling window trigger in Azure Data Factory?
Asked once · last asked Apr 14 · Concept · Mid-level
Show answer
A tumbling window trigger runs an Azure Data Factory pipeline for fixed, non-overlapping time windows. Each window has a start and end time, and the trigger keeps track of its state. It supports retries and dependencies between windows, which helps process time-based data reliably.
- Windows are fixed in size, consecutive, and do not overlap.
- Use window start and end times to select the data for each run.
- Supports retries, processing past windows, and dependencies on other tumbling window triggers.
What the interviewer checks: The interviewer is checking whether you understand time-window processing and how this differs from a simple schedule trigger.
What is the difference between Get Metadata and Lookup activity in Azure Data Factory?
Asked once · last asked Aug 19 · Concept · Junior
Show answer
Get Metadata reads information about a dataset, such as file size, column structure, or files in a folder. Lookup reads actual data from a file, table, or query result. Use Get Metadata to inspect a source and Lookup to fetch values that control later pipeline steps.
- Get Metadata can check whether a file or folder exists; supported fields depend on the source.
- Lookup returns the first row by default, or multiple rows when configured.
- Lookup returns at most 5,000 rows and fails if its output exceeds 4 MB.
- Neither activity is meant to copy data; use Copy activity for that.
What the interviewer checks: The interviewer checks whether you can choose between inspecting a source and reading its data.
If asked to redesign an Azure Data Factory pipeline as an Airflow DAG, what would you examine before translating it, and what would you avoid copying directly?
Asked once · last asked Sep 21 · Scenario · Senior
Show answer
I would examine dependencies, triggers, data volumes, compute needs, credentials, and failure handling before translating the pipeline. I would map logical work units to Airflow tasks, not copy every ADF activity directly. Airflow mainly coordinates work, so I would keep heavy data processing in external systems and redesign scheduling and retries for Airflow.
- Check linked services and integration runtimes for connectivity, authentication, and compute requirements.
- Map schedules and event triggers carefully; review Airflow data intervals, catchup, and backfills.
- Make tasks safe to rerun, and review retries, timeouts, and concurrency limits.
- Do not translate visual layouts or data flows directly; pass references, not large datasets, through XCom.
What the interviewer checks: The interviewer is checking whether you can preserve pipeline behavior while adapting to Airflow’s execution model.
How would you design an Azure Data Factory pipeline for a food delivery analytics project?
Asked once · last asked Oct 7 · Design · Mid-level
Show answer
I would use Azure Data Factory to load order, payment, and delivery data into a data lake. I would clean and join it using Mapping Data Flows or Databricks, then load reporting tables for Power BI. I would use incremental loads, data checks, and monitoring to keep reports reliable.
- Agree on metrics, such as delivery time, cancellation rate, and revenue, before designing tables.
- Use watermarks or change tracking for incremental loads, and handle late updates without creating duplicates.
- Add retries, failure alerts, and checks for missing orders or invalid timestamps.
- Protect customer data with managed identities, Key Vault for secrets, and limited access.
What the interviewer checks: The interviewer checks whether you can connect business metrics to a reliable, secure data pipeline.
What different kinds of triggers are available in Azure Data Factory?
Asked once · last asked Apr 14 · Concept · Junior
Show answer
Azure Data Factory has schedule, tumbling window, storage event, and custom event triggers. Schedule triggers run pipelines at set times, while tumbling window triggers process fixed, non-overlapping time windows and track their state. Storage event and custom event triggers run pipelines when matching events arrive through Azure Event Grid.
- Schedule triggers support recurring runs but do not track time-window completion.
- Tumbling window triggers support retries, dependencies, and processing past windows.
- Storage event triggers respond to blob creation or deletion in supported Azure storage accounts.
- Custom event triggers respond to application-defined events published to an Event Grid custom topic.
What the interviewer checks: The interviewer checks whether you can choose between time-based and event-based triggers for a pipeline.
What are the end-to-end steps for building a pipeline in Azure Data Factory?
Asked once · last asked Aug 10 · Tooling · Mid-level
Show answer
Define the source, destination, transformations, and schedule, then create a Data Factory. Configure access, linked services, datasets, and the required integration runtime. Build the pipeline with activities, dependencies, parameters, and error handling, then validate and debug it. Publish the changes, configure and publish a trigger if needed, and monitor runs for failures and performance.
- Linked services define connections; datasets describe the data that activities use.
- Choose an integration runtime that can reach your data sources and destinations.
- Test with sample data and check both activity results and destination data.
- Debug tests unpublished changes; publish before scheduled runs, and monitor failures with alerts.
What the interviewer checks: The interviewer checks whether you can take a pipeline from requirements through testing, deployment, and daily operation.
Other ways it was asked
- “Explain the end to end step to build a pipeline in Azure Data Factory.”
What format does the output of an Azure Data Factory Lookup activity use?
Asked once · last asked Aug 15 · Concept · Junior
Show answer
An Azure Data Factory Lookup activity returns JSON output. With “First row only” enabled, the result is in the `firstRow` object. When disabled, the rows are in the `value` array.
- “First row only” is enabled by default.
- Use `@activity('Lookup1').output.firstRow` to access the single row.
- Use `@activity('Lookup1').output.value` to access the array, for example in a ForEach activity.
What the interviewer checks: The interviewer checks whether you know the output format and how to access single-row and multiple-row results.
What are parameters and variables in Azure Data Factory?
Asked once · last asked Oct 7 · Concept · Junior
Show answer
Parameters pass values into pipelines, datasets, and linked services in Azure Data Factory. Their values stay fixed during a run. Variables hold temporary values inside a pipeline and can change during the run.
- Parameters make pipelines reusable, such as by passing a file path or load date.
- Pipeline variables support String, Bool, and Array types.
- Use Set Variable to assign a value and Append Variable to add an array item.
- Variables are shared across a pipeline run, so updating them in parallel loops is unsafe.
What the interviewer checks: The interviewer checks whether you can separate fixed inputs from changing pipeline state.
Azure Data Factory basics interview questions
Start with questions about what ADF does, how it fits with other Azure services, platform resources, and relevant Azure certifications. 5 questions.
Azure Data Factory activities and triggers interview questions
Explore pipeline activities, parameters, variables, triggers, reusable workflows, and how ADF orchestration experience transfers to Airflow. 10 questions.
Azure Data Factory Copy activity and incremental loading interview questions
These questions cover Copy activity, integration runtimes, source connections, file selection, metadata-driven ingestion, CDC, watermarks, and safe incremental reruns. 11 questions.
How do you configure connectivity from Azure Data Factory to an on-premises SQL Server, and what connection parameters are required?
Asked once · last asked Aug 19 · Tooling · Mid-level
Other ways it was asked
- “How do you establish connectivity and move data between Azure Data Factory and an on-premises SQL Server?”
- “How do you configure connectivity between Azure Data Factory and an on-premises SQL Server?”
How do you handle incremental data loading in Azure Data Factory, and how do you ensure the pipeline is rerunnable if it fails halfway?
Asked once · last asked May 9 · Tooling · Senior
When designing an Azure Data Factory pipeline to ingest high-volume P&C claims data, which method offers the most efficient incremental loading strategy: using a watermark column, performing a full daily refresh, leveraging change data capture (CDC), or using a time-based windowing function?
Asked once · last asked Jun 10 · Scenario · Senior
How can you design a generic Azure Data Factory pipeline to fetch data from 200 SQL tables and load it into ADLS?
Asked once · last asked Aug 15 · Design · Mid-level
How would you use Azure Data Factory to copy files from a temporary folder to a raw folder only when they contain records?
Asked once · last asked Aug 20 · Tooling · Mid-level
How would you use Azure Data Factory to read only the CSV files from a folder of thousands of mixed-format files in ADLS and merge them into a single CSV?
Asked once · last asked Aug 20 · Tooling · Mid-level
How would you use the wildcard file path in the Copy Data activity to handle multiple CSV files?
Asked once · last asked Aug 20 · Tooling · Mid-level
How would you design an ADF and Snowflake pipeline so corrections made in the source system after records are first created appear in the target without creating duplicates?
Asked once · last asked Sep 23 · Design · Senior
How would you load data from Azure Blob Storage into Databricks?
Asked once · last asked Oct 7 · Tooling · Mid-level
Which integration runtime would you use to copy data from Blob Storage to Databricks?
Asked once · last asked Oct 7 · Scenario · Mid-level
Azure Data Lake Storage and Event Hubs interview questions
Review storage locations, file formats, Delta tables, partitioning, Event Hubs sizing, message replay, and streaming ingestion. 8 questions.
Explain your strategy for partitioning in Azure Data Lake Storage to account for large-scale enterprise projects.
Asked once · last asked May 9 · Design · Senior
What determines the number of partitions configured for an Azure Event Hub?
Asked once · last asked Aug 7 · Concept · Mid-level
How can you read the same message from Azure Event Hubs again after consuming it?
Asked once · last asked Aug 7 · Scenario · Mid-level
How would you read data from Azure Event Hubs using Spark Structured Streaming and Kafka?
Asked once · last asked Aug 7 · Tooling · Mid-level
How is a Delta table different from data stored in Azure Data Lake Storage (ADLS)?
Asked once · last asked Oct 1 · Concept · Junior
Which file format would you choose for storing analytics data in Azure Blob Storage?
Asked once · last asked Oct 7 · Scenario · Mid-level
Can Databricks store table data itself, or do you need external storage such as ADLS?
Asked once · last asked Aug 26 · Concept · Mid-level
Other ways it was asked
- “Where are you creating the data lake—in Azure or in Databricks?”
- “Where are you creating the Data Lake, in Azure or in Databricks?”
- “Can you write data directly into Databricks tables without needing an external Azure storage layer like ADLS?”
Where is the data lake physically located when using Azure?
Asked once · last asked Aug 26 · Concept · Mid-level
Azure Data Factory data flows and PySpark interview questions
Work through schema mapping, type conversion, nested JSON, schema drift, and PySpark tasks for reading, filtering, joining, aggregating, and writing data. 11 questions.
Write a PySpark script to read CSV from ADLS, remove records with missing values in a specified column, group by another column to calculate an average, and write the results to a Delta Lake table in ADLS with schema enforcement and partitioning.
Asked once · last asked Aug 11 · Write code · Mid-level
Other ways it was asked
- “Can you write a PySpark script that reads a CSV file from Azure Data Lake Storage, filters out records where a specific column has no values, groups the data by a given column to calculate the average of another column, and writes the transformed data back to a Delta Lake table in ADLS with schema enforcement and partitioning?”
- “Can you write a PySpark script that reads a CSV file from Azure Data Lake Storage, filters out empty values, groups the data by a column to calculate an average, and writes the results to Delta Lake with schema enforcement and partitioning?”
Write a PySpark script to read a CSV from ADLS, calculate average sales per region, and save the result as Parquet, optimized for performance.
Asked once · last asked Aug 10 · Write code · Mid-level
Write a performance-optimized PySpark script to read large customer and transaction datasets in Parquet format from Azure Data Lake, inner join them on customer_id, calculate total transaction amounts per customer, and save the result as CSV in Azure Data Lake.
Asked once · last asked Aug 10 · Write code · Senior
Other ways it was asked
- “Write a PySpark script to join two large datasets stored in Parquet format in Azure Data Lake. The first data set contains customer details and the second contains transaction details. Perform an inner join on the customer_id column and then calculate the total transaction amount for each customer. Save the result back to Azure Data Lake in CSV format, ensuring optimal performance.”
How would you design an Azure Data Factory flow to consolidate large datasets from Azure SQL Database and Blob Storage, with differing schemas and data types, into one Azure Data Lake file while handling schema mapping, schema drift, scalability, and performance?
Asked once · last asked Aug 10 · Design · Senior
Other ways it was asked
- “You are tasked with creating a solution in Azure Data Factory to move data from multiple sources, including Azure SQL database and Blob storage, into a single consolidated file in Azure Data Lake. The data sources vary in schema and data types. How would you design the data flow to handle schema mapping, ensure scalability, and optimize performance for processing large datasets? Additionally, how would you handle schema drift during this process?”
How do you map different schemas and data types in Azure Data Factory while handling invalid values?
Asked once · last asked Aug 10 · Scenario · Mid-level
Other ways it was asked
- “How would you map columns between datasets with completely different structures in Azure Data Factory, and ensure data type compatibility during transformations?”
- “To simplify this further, if you have two datasets in Azure Data Factory, one containing sales data and another containing customer demographic data, both with different schemas, how would you map the columns between these datasets to create a unified schema while ensuring correct data types? Please provide an example of handling a mismatched data type during this process.”
- “How would you map columns with different data types, such as a string column to a numeric column, in Azure Data Factory, and prevent invalid data from disrupting the pipeline?”
How would you use PySpark to process millions of Azure Data Lake log entries with timestamp and event_type columns into daily event counts, pre-aggregate in memory despite memory limits, handle event_type skew during grouping and partitioning, and write date-partitioned Parquet back to Azure Data Lake?
Asked once · last asked Aug 10 · Scenario · Senior
Other ways it was asked
- “Imagine you are tasked with creating a PySpark based solution to process a dataset containing millions of log entries stored in Azure Data Lake. The data set includes a timestamp column and an event_type column. Your goal is to generate a daily summary of event counts grouped by event_type and save the results as a partitioned parquet file, partitioned by date, back to Azure Data Lake. How would you modify your PySpark script to ensure that the event_type column is pre-aggregated in memory before writing to the parquet file, considering scenarios where the dataset could exceed the available memory? Additionally, how would you handle skewed data in the event_type column to avoid bottlenecks during grouping and partitioning?”
Write a PySpark script to read a dataset containing event_type and timestamp columns from Azure Data Lake, filter for specific event_type values such as click and purchase, count occurrences of each event_type, and save the results as CSV in Azure Data Lake.
Asked once · last asked Aug 10 · Write code · Mid-level
Other ways it was asked
- “Let's simplify this. Write a PySpark script to process a dataset stored in Azure Data Lake. The dataset includes event_type and timestamp columns. Your task is to filter only the rows with specific event_type values, e.g. click, purchase, and then count the occurrences of each event_type. Save the results as a CSV file back to Azure Data Lake.”
How would you design an Azure Data Factory pipeline to ingest multiple JSON files from Azure Blob Storage, parse nested structures, extract selected fields, and flatten them into an Azure SQL Database table while handling schema drift, performance, and data consistency?
Asked once · last asked Aug 10 · Design · Senior
Other ways it was asked
- “You are working on an Azure Data Factory pipeline that ingests data from multiple JSON files stored in Azure Blob Storage. The pipeline needs to parse nested JSON structures, extract specific fields, and transform the data into a flattened schema before writing it into an Azure SQL database table. How would you design this pipeline? Include details on handling schema drift, optimizing performance, and ensuring data consistency during the process.”
How would you configure an Azure Data Factory mapping data flow to extract nested JSON fields while preserving records when some fields are missing or null?
Asked once · last asked Aug 10 · Tooling · Mid-level
Other ways it was asked
- “How would you configure the explode_outer operation in an Azure Data Factory mapping data flow to extract nested JSON fields, including when some fields are missing or null?”
Write a PySpark script to load a single JSON file from an Azure Data Lake storage, extract a specific field such as customer_id, and save the results as a partitioned Delta table by customer_id, handling missing or null customer_id values.
Asked once · last asked Aug 22 · Write code · Mid-level
Azure Synapse Analytics interview questions
Synapse questions focus on compute, distributions, dedicated and serverless SQL pools, workload management, schema changes, and query performance. 7 questions.
What is Azure Synapse Analytics, and how are its compute resources and data distributions organized?
Asked once · last asked Feb 4 · Concept · Senior
Other ways it was asked
- “What is Azure Synapse Analytics, and how does its compute work?”
- “In Azure Synapse Analytics, how many distributions are there, how is compute mapped to them, and what compute configurations are available?”
- “In Azure Synapse Analytics, does each distribution have its own compute resource, or can one compute resource access data across multiple distributions?”
Do Snowflake and Azure Synapse Analytics use the same compute and data distribution model?
Asked once · last asked Feb 4 · Concept · Mid-level
If a Synapse dedicated SQL pool query is running slowly on a multi-terabyte table, what steps would you take to optimize it?
Asked once · last asked May 9 · Scenario · Senior
How does workload management work in Azure Synapse Analytics?
Asked once · last asked Aug 7 · Concept · Mid-level
How would you design a solution to load and query petabytes of data stored in Azure Data Lake using Azure Synapse Analytics, including data structure, query performance, and schema evolution?
Asked once · last asked Aug 11 · Design · Senior
How would you structure data and use serverless SQL pools to ensure cost-effective query performance in Azure Synapse Analytics?
Asked once · last asked Aug 11 · Design · Senior
How would you handle frequent schema evolution in Azure Synapse Analytics without breaking existing queries or pipelines, and what features or techniques would you use to ensure seamless schema evolution?
Asked once · last asked Aug 11 · Scenario · Senior
Microsoft Fabric and Power BI interview questions
Fabric and Power BI questions cover pipelines, notebooks, warehouses, OneLake, SQL endpoints, semantic models, reporting modes, and capacity planning. 22 questions.
Why would you use user-defined functions in Microsoft Fabric, and how would you use them?
Asked once · last asked Sep 15 · Scenario · Mid-level
Other ways it was asked
- “What are user-defined functions in Fabric?”
- “How have you used user-defined functions in Microsoft Fabric? Give an example.”
Can you rename a column in a Microsoft Fabric Warehouse using the SQL endpoint?
Asked once · last asked Sep 16 · Concept · Mid-level
Other ways it was asked
- “How can you rename a column programmatically in a Microsoft Fabric table?”
How would you estimate the Microsoft Fabric capacity a company needs?
Asked once · last asked May 20 · Scenario · Senior
Why might records written through a Microsoft Fabric notebook not appear immediately through the Lakehouse SQL endpoint, and how would you handle it?
Asked once · last asked Sep 7 · Scenario · Senior
Can you explain the difference between DirectQuery and Direct Lake in Microsoft Fabric?
Asked once · last asked Sep 7 · Concept · Mid-level
How do you choose between PySpark and SQL when creating notebooks in Microsoft Fabric?
Asked once · last asked Sep 9 · Concept · Mid-level
How would you build a Microsoft Fabric pipeline to ingest, transform, and load data from three sources into one Lakehouse?
Asked once · last asked Sep 9 · Design · Mid-level
Would you use parent and child pipelines in Microsoft Fabric, and how would you implement them?
Asked once · last asked Sep 9 · Design · Mid-level
What is the best way to build multiple reports in Microsoft Fabric on a shared semantic model?
Asked once · last asked Sep 9 · Design · Mid-level
In Microsoft Fabric, can you use a snowflake schema in Power BI while maintaining report performance, and how do modeling limitations affect that choice?
Asked once · last asked Sep 9 · Scenario · Senior
How would you use Microsoft Fabric to work with data from Fabric, Google Cloud Platform, and SQL sources, and optimize the solution?
Asked once · last asked Sep 9 · Design · Senior
How do you make a reusable PySpark function available across Microsoft Fabric notebooks?
Asked once · last asked Sep 15 · Tooling · Mid-level
Show 9 more
Why use notebooks in Microsoft Fabric, and when would you use them with a data warehouse or OneLake?
Asked once · last asked Sep 16 · Concept · Mid-level
Can you rename a table in OneLake using the SQL endpoint?
Asked once · last asked Sep 16 · Concept · Mid-level
What common issues do you encounter in Microsoft Fabric data pipelines, and how do you address them?
Asked once · last asked Sep 16 · Scenario · Mid-level
If you write data to OneLake instead of ADLS, how would that affect your technology and data-format choices?
Asked once · last asked Sep 16 · Scenario · Senior
How do you integrate Power BI with Microsoft Fabric?
Asked once · last asked Sep 18 · Tooling · Mid-level
How would you design a data solution using Microsoft Fabric, including its components, integrations, and workflow?
Asked once · last asked Sep 24 · Design · Mid-level
Other ways it was asked
- “Can you describe a project where you used Microsoft Fabric, including the components, integrations, and workflow involved?”
Which Microsoft Fabric item provides relational data capabilities and supports querying?
Asked once · last asked Sep 16 · Concept · Junior
Which Microsoft Fabric capability can schedule data ingestion workflows?
Asked once · last asked Sep 16 · Concept · Junior
Why is Data Factory used in Microsoft Fabric?
Asked once · last asked Sep 18 · Concept · Junior
Azure Data Factory monitoring and performance interview questions
Troubleshooting questions cover pipeline performance, alerts, incomplete batches, reconciliation, report discrepancies, and recovery documentation. 10 questions.
Can you tell me about a time you had to troubleshoot a data pipeline or refresh problem in the Azure environment, and what you did?
Asked once · last asked Apr 24 · Behavioral · Mid-level
Imagine you are tasked with optimizing an Azure Data Factory pipeline that processes a daily large volume of data. How would you approach this?
Asked once · last asked Aug 10 · Scenario · Senior
Other ways it was asked
- “How would you approach identifying the bottlenecks and optimizing an Azure Data Factory pipeline that takes five hours to complete?”
When joining multiple large tables or when the smaller table cannot fit in memory, how would you improve join efficiency and reduce shuffling without broadcast joins, while also identifying and addressing indexing issues in Azure SQL?
Asked once · last asked Aug 10 · Scenario · Senior
Other ways it was asked
- “When a smaller table cannot fit in memory or you are joining multiple large tables, what alternatives to broadcast joins can reduce data shuffling and improve join efficiency? How would you identify and address indexing issues in Azure SQL?”
An ADF activity returns success, but what would tell you that the extraction stopped too early? How would you change the extraction logic and tests so the same problem is detected before the data reaches Power BI?
Asked once · last asked Sep 28 · Scenario · Senior
If row counts match after an Oracle-to-Fabric migration but a report shows a different monthly total in Fabric, how would you isolate whether the difference comes from source mapping, transformation logic, or the report itself?
Asked once · last asked Sep 29 · Scenario · Senior
If a new engineer has to support an incremental ADF-to-Snowflake pipeline while you are away, what would you document so they can detect a partial batch, rerun it safely, and prove the recovered data is complete?
Asked once · last asked Sep 29 · Scenario · Senior
If an ADF activity succeeds and raw data is present in Snowflake but a curated table has not changed for two days, how would you distinguish a legitimate no-change period from a failed transformation or stuck incremental checkpoint?
Asked once · last asked Sep 29 · Scenario · Senior
What differences between Snowflake and Microsoft Fabric can affect reporting results, and how would you check whether a platform difference changed the output?
Asked once · last asked Oct 1 · Concept · Senior
How would you monitor Azure Data Factory pipelines and notify business stakeholders when a pipeline fails?
Asked once · last asked Oct 7 · Tooling · Mid-level
Azure data architecture and migration interview questions
Think through end-to-end data designs, platform comparisons, cross-cloud migrations, domain modeling, security, governance, and ownership. 19 questions.
How would you design a schema and choose an Azure data store—Azure SQL, Cosmos DB, or Azure Data Lake—for high-volume invoice processing and real-time fraud detection, ensuring encryption at rest, row-level security, and scalable reads and writes?
Asked once · last asked Mar 9 · Design · Senior
Other ways it was asked
- “How would you design a database schema and choose an Azure data store, such as Azure SQL, Cosmos DB, or Azure Data Lake, for high-volume invoice processing and real-time fraud detection while ensuring encryption at rest, row-level security, and scalable read and write performance?”
How would you approach a complex data modeling challenge in the property and casualty (P&C) insurance domain, and which Azure services would you use?
Asked once · last asked Jun 10 · Behavioral · Senior
Other ways it was asked
- “Describe a complex data modeling challenge you faced in the property and casualty (P&C) insurance domain, including your approach to resolving it and the specific Azure services you used.”
For long-term data governance success on a new Azure P&C insurance platform, which aspects are paramount: clear cross-business data ownership and accountability, automated data quality monitoring and remediation, one centralized Azure data lake, least-privilege access controls, or comprehensive metadata management and cataloging?
Asked once · last asked Jun 10 · Scenario · Senior
Other ways it was asked
- “When establishing data governance for a new Azure data platform in P&C insurance, which aspects are paramount for long-term success: clear data ownership and accountability across business units, automated data quality monitoring and remediation, storing all data in one centralized Azure data lake, least-privilege security and access controls, or comprehensive metadata management and data cataloging?”
What target-state architecture would you propose for modernizing an on-premises Hadoop platform to Databricks on Azure?
Asked once · last asked Jul 16 · Design · Senior
How would you roll out a Spark-based data platform using Microsoft Fabric, Azure HDInsight, or Azure Synapse?
Asked once · last asked Jul 28 · Scenario · Senior
How would you approach data governance with Microsoft Purview for lineage and classification?
Asked once · last asked Jul 28 · Scenario · Senior
How can Databricks and Airflow be integrated with Azure data services?
Asked once · last asked Aug 7 · Design · Mid-level
Why use Databricks for transformations when Azure has native tools to process data in the lake?
Asked once · last asked Aug 26 · Concept · Mid-level
Why use Databricks to create and manage tables instead of using native Azure services like Azure SQL?
Asked once · last asked Aug 26 · Concept · Mid-level
Why would you choose Databricks over Synapse for creating tables and managing the middle layer of your architecture?
Asked once · last asked Aug 26 · Concept · Mid-level
Suppose we needed to move hundreds of terabytes of data between platforms, for example from Snowflake on AWS to Microsoft Fabric on Azure. How would you design the migration while controlling cost, maintaining security, and validating the data?
Asked once · last asked Aug 31 · Design · Senior
When would you use Azure Data Factory or Microsoft Fabric rather than Databricks or Spark for a large historical migration from Snowflake into Fabric Delta tables?
Asked once · last asked Aug 31 · Tooling · Senior
Show 6 more
How would you design an Azure data pipeline for a daily 10 GB ingestion use case?
Asked once · last asked Sep 1 · Design · Mid-level
How would you migrate nonstandardized Snowflake data on AWS to Azure using CDC while decoupling reporting from ERP systems?
Asked once · last asked Sep 16 · Design · Senior
Other ways it was asked
- “How would you design a solution to migrate nonstandardized Snowflake data on AWS to Azure using CDC Replicate and a medallion architecture?”
How would you approach the migration of data from Snowflake on AWS to Azure in phases?
Asked once · last asked Sep 16 · Design · Senior
How would you decide between OneLake and Databricks Lakehouse?
Asked once · last asked Sep 16 · Scenario · Senior
How would you work with teams to define schemas before moving data into Microsoft Fabric?
Asked once · last asked Sep 18 · Scenario · Mid-level
During a Microsoft Fabric migration, how would you investigate an undocumented transformation rule that changes an important report before deciding whether to reproduce or remove it?
Asked once · last asked Sep 21 · Scenario · Senior
Related interview questions
- Data Engineer interview questions
- Data Analyst interview questions
- System Administrator interview questions
- DevOps Engineer interview questions
- Cloud Engineer interview questions
- SQL interview questions
- MongoDB and NoSQL interview questions
- ETL and Data Pipelines interview questions
- Data Modeling interview questions
- Spark, PySpark and Databricks interview questions
- Snowflake interview questions
- Power BI interview questions
- Excel interview questions
How this list is made
Every question here was asked by a real interviewer in a real job interview that a candidate ran MeetAssist in. After the call, the interviewer's questions are pulled from the transcript. Practice sessions, tests, exams and anything that was not a job interview are dropped.
Questions are rewritten so they make sense on their own: names of people, employers and clients are removed, and speech-to-text mistakes are fixed. Small talk, logistics and questions about one candidate's own résumé are left out. Rephrasings of the same question are merged, and the count is the number of different interviews it came up in.
The interviews come from candidates who use MeetAssist, so they lean towards software, data and AI roles: the shares describe these interviews, not every Azure Data Factory interview. Answers to the most asked questions are written and checked by an editor; a "common mistake" is shown only when reviews of real answers in these interviews show it, with the count.
42 interviews, 17 candidates, Feb 2026 to Oct 2026. Updated every month with new interviews.
Or add the Chrome extension