AWS Data Engineer Associate Sample Questions With Full Explanations
Here are 8 multiple-choice questions designed to mirror the style, difficulty, and domain coverage of the AWS Data Engineer Associate (DEA-C01) exam. The questions span data ingestion, storage, processing, orchestration, and governance.
Instructions: Choose the single best answer for each question. The questions are designed to test not just recall, but your ability to apply AWS services to real-world data engineering scenarios. We recommend pairing this with AWS Machine Learning Engineer Sample Questions With Full Explanations.
Question 1
A data engineering team needs to load streaming JSON data from Amazon Kinesis Data Streams into Amazon S3, with files partitioned by a timestamp derived from the event payload (not the ingestion time). They want to minimize operational overhead and avoid managing Spark clusters. Which AWS service is the most cost-effective and efficient choice for this task?
A. AWS Glue Streaming ETL job with a custom timestamp extractor
B. Amazon Kinesis Data Firehose with dynamic partitioning
C. Amazon Managed Streaming for Apache Kafka (MSK) with Kafka Connect
D. AWS Lambda function triggered by Kinesis, writing directly to S3
Answer: B
Explanation: Kinesis Data Firehose natively supports dynamic partitioning, allowing you to extract timestamps from the payload and write to S3 prefixes without managing servers. While AWS Glue can do this, it requires provisioning and managing a Spark cluster, making it less cost-effective for simple transformations. MSK adds more complexity than needed, and Lambda is not ideal for high-throughput, continuous streaming loads due to concurrency limits and potential data loss.
Question 2
A company has a large Amazon Redshift cluster used for BI reporting. They need to run a complex transformation that joins data from Amazon S3 (raw logs) with data already in Redshift, but they want to ensure the heavy lifting of the join does not degrade performance for existing BI queries. What is the most appropriate strategy?
A. Perform the join directly in Redshift using a CREATE TABLE AS (CTAS) statement.
B. Use Amazon Redshift Spectrum to query the S3 data and join it with local tables, pushing down the heavy processing to the Redshift cluster.
C. Use AWS Glue to read the S3 data and the Redshift table, perform the join in Glue, and write the results back to Redshift.
D. Load the S3 data into a temporary Redshift table and run the join, then drop the temporary table.
Answer: C
Explanation: Offloading the join to AWS Glue (serverless Spark) prevents the processing from consuming Redshift's compute resources, ensuring BI query performance is unaffected. While Redshift Spectrum pushes down filters, the join itself between Spectrum and local tables often executes on the Redshift cluster, which can still cause contention. CTAS and temporary tables directly consume Redshift compute, which is exactly what the team wants to avoid.
Question 3
A data pipeline uses AWS Glue ETL jobs to process data every hour. The jobs read from a JDBC connection to a legacy database. The team has noticed that the jobs are slow and are concerned about the load on the source database. Which feature should they enable to improve performance and reduce the load?
A. Enable Glue job bookmarking.
B. Increase the number of Glue Data Processing Units (DPUs) for the job.
C. Use a JDBC connection with a fetchSize parameter set to a larger value.
D. Enable S3 connection for the JDBC source.
Answer: C
Explanation: Setting a larger fetchSize allows the Glue job to retrieve more rows per network round-trip, reducing the number of queries sent to the database and significantly speeding up data transfer. Job bookmarks help with incremental processing, not the raw throughput of a full load. Increasing DPUs adds more Spark executors, which can actually increase the load on the source DB if they all query simultaneously. S3 connection is irrelevant for a JDBC source.
Question 4
A data engineer is designing a data lake on Amazon S3. They need to store data in a way that allows multiple analytics tools (Amazon Athena, Amazon Redshift Spectrum, and Amazon EMR) to query it directly without transformation. They also need to enforce schema evolution. Which approach best meets these requirements?
A. Store data as JSON in S3 and use an AWS Glue Data Catalog table with a schema that is manually updated.
B. Store data as Parquet in S3 and register the table in the AWS Glue Data Catalog using a crawler. You can explore this further in Google Professional Data Engineer Sample Questions With Full Explanations.
C. Store data as CSV in S3 and use Athena's CREATE TABLE AS (CTAS) to generate a schema.
D. Store data as Avro in S3 and use Amazon DynamoDB to manage the schema.
Answer: B
Explanation: Parquet is a columnar format optimized for analytics, and the AWS Glue Data Catalog acts as a central metadata repository that all the mentioned services (Athena, Spectrum, EMR) natively integrate with. Using a crawler helps automate schema discovery and can handle some evolution. While JSON is flexible, it is less performant for large analytics queries. CTAS generates data, not a schema for existing data. DynamoDB is not a schema registry for S3 data.
Question 5
A company is building a near-real-time analytics dashboard. They have an application writing events to Amazon Kinesis Data Streams. They need to aggregate these events by a key (e.g., user ID) every 1 minute and load the results into Amazon OpenSearch Service. Which AWS service is specifically designed to perform this stateful stream processing with exactly-once semantics?
A. Amazon Kinesis Data Firehose
B. AWS Lambda with a Kinesis trigger
C. Amazon Kinesis Data Analytics (now part of Managed Service for Apache Flink)
D. AWS Glue crawler
Answer: C
Explanation: Kinesis Data Analytics (Managed Flink) uses Apache Flink, which provides stateful stream processing capabilities—like windowed aggregations (1-minute tumbling windows) and exactly-once processing—that are required for this use case. Kinesis Data Firehose is for streaming data into destinations without complex transformations (no windowed state). Lambda is stateless and has execution time limits, making it poor for stateful aggregations. A Glue crawler is for metadata discovery, not stream processing.
Question 6
A data engineer needs to orchestrate a complex ETL pipeline that includes an AWS Glue job, a Redshift SQL query, and an external API call. The pipeline must run based on a schedule and also be triggerable by new data arriving in S3. What is the best tool to build this workflow?
A. AWS Step Functions
B. Amazon EventBridge Scheduler
C. AWS Glue Workflows
D. Amazon Managed Workflows for Apache Airflow (MWAA)
Answer: A
Explanation: AWS Step Functions is a serverless orchestration service that excels at coordinating multiple AWS services (Glue, Redshift Data API, Lambda for API calls) into a state machine. It supports both scheduled triggers (via EventBridge) and event-driven triggers (via S3 event notifications). While MWAA is powerful, it is more complex to manage (requires VPC setup, etc.). Glue Workflows are limited to orchestrating Glue jobs and crawlers, not external API calls. EventBridge Scheduler only handles the trigger, not the orchestration logic.
Question 7
A team is migrating an on-premises Apache Hadoop cluster to AWS. They want to minimize costs by only paying for compute when they are processing data. They have a mix of scheduled batch jobs and ad-hoc interactive queries. Which Amazon EMR architecture is most suitable?
A. A long-running, persistent cluster with a core node group and task node group.
B. A transient cluster that is started for each scheduled job and terminated after completion.
C. A persistent cluster using only Spot Instances for all nodes.
D. A serverless cluster using Amazon EMR Serverless.
Answer: D Before moving on, consider reading AWS AI Practitioner Sample Questions With Full Explanations.
Explanation: Amazon EMR Serverless is the best fit for mixed batch and interactive workloads, as it automatically provisions and scales compute in response to demand, and you only pay for the vCPU and memory consumed by your jobs. This eliminates the need to manage clusters and avoids paying for idle time. Transient clusters are good for batch but require startup time for interactive queries. Persistent clusters incur costs even when idle, and using only Spot Instances risks job interruptions.
Question 8
A data engineer is troubleshooting an AWS Glue ETL job that is failing. The error logs are written to Amazon CloudWatch, but the engineer wants to see the Spark driver and executor logs aggregated in one place for easier debugging. What should they configure to achieve this?
A. Enable AWS CloudTrail for the Glue job.
B. Configure the Glue job to write logs to an Amazon S3 bucket.
C. Enable "Continuous logging" and "Spark UI" for the Glue job.
D. Set the Glue job's --enable-metrics parameter to true.
Answer: C
Explanation: AWS Glue's "Continuous logging" feature streams logs from the Spark driver and executors to CloudWatch in near-real-time, while the "Spark UI" feature provides a web UI for viewing job progress and metrics. Together, they offer a comprehensive view for debugging. CloudTrail records API calls, not application logs. Writing to S3 is possible but not the standard built-in method for log aggregation. --enable-metrics only provides CloudWatch metrics, not detailed log lines.
Question 9
A company stores sensitive customer data in an Amazon S3 data lake. A Data Engineer needs to ensure that this data is encrypted at rest and that the encryption keys are rotated automatically every year. Which configuration meets these requirements with the least operational effort?
A. Server-Side Encryption with Customer-Provided Keys (SSE-C).
B. Server-Side Encryption with S3-Managed Keys (SSE-S3).
C. Server-Side Encryption with AWS KMS-Managed Keys (SSE-KMS) with automatic key rotation enabled.
D. Client-Side Encryption using the AWS Encryption SDK.
Answer: C
Explanation: SSE-KMS provides envelope encryption and integrates with AWS KMS, where you can enable automatic key rotation (yearly) for the Customer Master Key (CMK). This ensures data is encrypted at rest with minimal operational overhead. SSE-S3 encrypts with keys managed by S3, but does not offer customer-controlled rotation. SSE-C requires you to manage and provide your own keys, adding significant operational burden. Client-side encryption requires you to manage the entire encryption lifecycle within your application code.
Question 10
A Data Engineer needs to perform a one-time migration of 50 TB of historical data from an on-premises HDFS cluster to Amazon S3. The network bandwidth between the data center and AWS is limited to 100 Mbps. The migration must be completed within 5 days. What is the most reliable and cost-effective solution?
A. Use AWS DataSync to transfer the data over the network.
B. Use the AWS CLI aws s3 sync command to copy the data.
C. Order an AWS Snowball Edge Storage Optimized device.
D. Set up a VPN connection and use an Amazon S3 multipart upload script.
Answer: C
Explanation: At 100 Mbps, transferring 50 TB over the network would take approximately 46 days (50 TB / (100 Mbps / 8)), which far exceeds the 5-day deadline. AWS Snowball Edge is a physical device designed for petabyte-scale data transfer, bypassing the network bottleneck entirely. DataSync and aws s3 sync are network-based and would be too slow. A VPN with multipart upload is an even more complex and slower version of the network-based options.
These questions represent the core competencies required for the AWS Data Engineer Associate certification: designing data stores, orchestrating pipelines, and implementing data processing and analysis. Remember to focus on the "why" behind each answer to truly understand the AWS Well-Architected principles as they apply to data engineering.
To structure your study time around those competencies, a free AWS Certified Data Engineer - Associate study plan can give you a week-by-week schedule tailored to your available hours.
Turn passive knowledge into exam-passing recall with AWS Certified Data Engineer - Associate practice exam questions.
A free AWS Certified Data Engineer - Associate readiness check gives you a quick way to see which topics need more review before the exam.