Career upgrade: Learn practical AI skills for better jobs and higher pay.
Level up
PracticeVideosBlogFlashcardsEspañol
All Practice Exams

100+ Free Cloudera Data Engineer Practice Questions

Pass your Cloudera Certified Data Engineer (CDP-3002) exam on the first try — instant access, no signup required.

✓ No registration✓ No credit card✓ No hidden fees✓ Start practicing immediately
100+ Questions
100% Free
1 / 100
Question 1
Score: 0/0

A Spark application needs to perform a sort-merge join between two large tables. Which prerequisite must be met before the join can execute?

A
B
C
D
to track
2026 Statistics

Key Facts: Cloudera Data Engineer Exam

50

Exam Questions

Cloudera CDP-3002 exam guide

90 min

Exam Duration

Cloudera CDP-3002 exam guide

55%

Passing Score

Cloudera CDP-3002 exam guide

$330

Exam Fee

Cloudera Education Store (2026)

2 years

Certification Validity

Cloudera Certification FAQ

48%

Spark Domain Weight

Cloudera CDP-3002 exam guide

Cloudera's CDP-3002 Data Engineer exam is a 50-question, 90-minute online proctored test delivered via Questionmark with a 55% passing score. The $330 exam emphasizes Spark on Kubernetes (48%), performance tuning (22%), Airflow (10%), deployment (10%), and Iceberg (10%). Certification is valid for 2 years. The old CCP Data Engineer (DE575) has been discontinued in favor of this CDP-era multiple-choice format.

Sample Cloudera Data Engineer Practice Questions

Try these sample questions to test your Cloudera Data Engineer exam readiness. Each question includes a detailed explanation. Start the interactive quiz above for the full 100+ question experience with AI tutoring.

1In Cloudera Data Engineering (CDE), Spark applications run on Kubernetes. Which component is responsible for requesting executor pods from the Kubernetes cluster?
A.The CDE virtual cluster scheduler
B.The Spark Driver pod
C.The YARN ResourceManager
D.Apache Airflow's KubernetesPodOperator
Explanation: In Spark on Kubernetes, the Spark Driver pod is responsible for requesting executor pods from the Kubernetes API server. The driver coordinates the application lifecycle, including requesting resources (executor pods) and distributing tasks across them.
2Which Spark method creates a new DataFrame by selecting specific columns from an existing DataFrame?
A.filter()
B.select()
C.groupBy()
D.join()
Explanation: The select() method in Spark is used to choose specific columns from a DataFrame, returning a new DataFrame containing only the selected columns. It is one of the most fundamental DataFrame transformation operations.
3A Spark job running on CDE shows uneven task durations across executors. One executor takes significantly longer than the others. What is the most likely cause?
A.Insufficient driver memory
B.Data skew in the partition being processed
C.Too many Airflow DAG retries
D.The Iceberg table has too many snapshots
Explanation: Data skew occurs when one partition holds significantly more data than others, causing the executor processing that partition to take much longer. This is a common issue in distributed processing and can be addressed with techniques like salting keys or using Spark's adaptive query execution (AQE).
4What is the primary purpose of the spark.kubernetes.container.image configuration property?
A.To specify the Docker image used for driver and executor pods
B.To define the maximum memory for each Kubernetes node
C.To set the container restart policy on failure
D.To configure the Kubernetes namespace for Spark pods
Explanation: The spark.kubernetes.container.image property specifies the Docker container image that Kubernetes will use when creating the Spark driver and executor pods. This image must contain the Spark runtime and any application dependencies.
5Which DataFrame operation triggers the actual execution of a Spark job?
A.withColumn()
B.filter()
C.count()
D.select()
Explanation: count() is an action in Spark that triggers the execution of the entire DAG of transformations. Spark uses lazy evaluation — transformations like filter(), select(), and withColumn() only build a logical plan. Actions like count(), collect(), and show() cause the plan to be executed.
6In Spark's Catalyst optimizer, what is the role of the logical plan?
A.It determines the physical storage format for output files
B.It represents the sequence of transformations without specifying how they will be executed
C.It schedules tasks across executors in the cluster
D.It manages memory allocation between storage and execution regions
Explanation: The logical plan in Spark's Catalyst optimizer represents the abstract sequence of transformations (filters, joins, projections) defined by the user without specifying execution details. The optimizer then converts the logical plan into an optimized physical plan that determines how transformations are actually executed.
7When integrating Spark with Hive in CDE, which catalog implementation allows Spark to read and write Hive-managed tables?
A.HiveMetastoreClient
B.SparkCatalog with Hive support enabled
C.HiveCatalog (Hive Metastore)
D.JDBCCatalog
Explanation: HiveCatalog connects Spark to the Hive Metastore, enabling Spark to discover, read, and write Hive-managed tables. In CDE, enabling Hive support through the HiveCatalog ensures that Spark applications can seamlessly access Hive tables and their metadata.
8What is the default persistence level when calling cache() on a Spark DataFrame?
A.MEMORY_ONLY
B.MEMORY_AND_DISK
C.DISK_ONLY
D.MEMORY_ONLY_SER
Explanation: When cache() is called on a DataFrame (or RDD), Spark uses the MEMORY_ONLY storage level by default for RDDs, and MEMORY_AND_DISK for DataFrames. However, cache() on a DataFrame is an alias for persist(StorageLevel.MEMORY_AND_DISK) in the Dataset API. For the purpose of the CDP exam, MEMORY_ONLY is the classic default for RDD cache(), while DataFrames default to MEMORY_AND_DISK.
9An Airflow DAG needs to extract only records that have changed since the last successful run. Which approach best implements incremental extraction?
A.Use a full table scan on every run and deduplicate afterward
B.Track a high-water mark (e.g., last modified timestamp) and filter source records newer than that value
C.Delete all target data before each run and reload everything
D.Schedule the DAG to run only once per month to reduce data volume
Explanation: Incremental extraction uses a high-water mark — typically a timestamp or monotonically increasing ID — to track the last successfully extracted record. Each DAG run queries the source for records newer than the stored high-water mark, significantly reducing data transfer and processing time compared to full loads.
10Which Spark configuration parameter controls the number of partitions produced after a shuffle operation?
A.spark.executor.instances
B.spark.sql.shuffle.partitions
C.spark.driver.memory
D.spark.kubernetes.executor.request.cores
Explanation: spark.sql.shuffle.partitions controls the number of partitions used when shuffling data for joins, aggregations, and other wide transformations in Spark SQL. The default value is 200, but it should be tuned based on data size and cluster resources to avoid too many small tasks or too few large ones.

About the Cloudera Data Engineer Exam

The Cloudera Data Engineer (CDP-3002) certification validates your ability to design, develop, and optimize data workflows using Spark on Kubernetes, Apache Airflow orchestration, and Apache Iceberg on the Cloudera Data Platform. The exam covers performance tuning, CDE deployment, and Hive integration.

Questions

50 scored questions

Time Limit

90 minutes

Passing Score

55%

Exam Fee

$330 (Cloudera)

Cloudera Data Engineer Exam Content Outline

48%

Spark

Spark fundamentals on Kubernetes, DataFrames, distributed processing, Hive-Spark integration, and distributed persistence formats

22%

Performance Tuning

Spark tuning tools, Catalyst optimizer, EXPLAIN plans, schema inference, join optimization, caching strategies, and partitioned/bucketed tables

10%

Airflow

Incremental extraction, ETL pipeline scheduling, quality check automation, and DAG configuration in Apache Airflow

10%

Deployment

CDE API and CLI usage, virtual clusters, job definitions, resource management, and Data Engineering Service operations

10%

Iceberg

Apache Iceberg fundamentals including ACID transactions, time travel, snapshot management, schema evolution, and partition evolution

How to Pass the Cloudera Data Engineer Exam

What You Need to Know

  • Passing score: 55%
  • Exam length: 50 questions
  • Time limit: 90 minutes
  • Exam fee: $330

Keys to Passing

  • Complete 500+ practice questions
  • Score 80%+ consistently before scheduling
  • Focus on highest-weighted sections
  • Use our AI tutor for tough concepts

Cloudera Data Engineer Study Tips from Top Performers

1Prioritize Spark topics — they account for nearly half (48%) of the exam questions
2Practice reading and interpreting Spark EXPLAIN plans to identify join strategies and predicate pushdown
3Build hands-on experience with Airflow DAGs including task dependencies, scheduling, and XComs
4Understand Iceberg fundamentals: snapshots, time travel, schema evolution, and partition evolution
5Know the difference between coalesce() and repartition() and when to use each for performance tuning

Frequently Asked Questions

How many questions are on the Cloudera Data Engineer CDP-3002 exam?

The CDP-3002 exam contains 50 multiple-choice and multi-select questions. You have 90 minutes to complete them, giving approximately 1.8 minutes per question.

What is the passing score for the Cloudera CDP-3002 exam?

Cloudera lists a 55% passing score for the CDP-3002 Data Engineer exam. This means you need to answer at least 28 out of 50 questions correctly.

How much does the Cloudera Data Engineer certification cost?

The CDP-3002 exam costs $330 USD. Each retake attempt also costs $330 with a mandatory 7-day waiting period between attempts.

How is the Cloudera CDP-3002 exam delivered?

The exam is delivered online via Questionmark with live proctoring. You need a webcam, stable internet connection, and a quiet, private room. No external resources, notes, or websites are permitted.

How long is Cloudera CDP-3002 certification valid?

CDP Certification Program certifications are valid for 2 years from the achievement date. You must retake the exam to maintain your certification status after it expires.

What topics should I focus on for the CDP-3002 exam?

Spark is the dominant domain at 48% of the exam. Focus on Spark on Kubernetes, DataFrames, and Hive integration. Performance tuning (22%) is the second priority, covering EXPLAIN plans, join optimization, and partitioning. Airflow, deployment, and Iceberg each account for 10%.