1
GCP Professional Data Engineer
Certification Exam Actual Exam
2026/2027 with Detailed Rationales |
Complete Exam-Style Questions | Pass
Guaranteed – A+ Graded
TABLE OF CONTENTS
Section 1 | Designing Data Processing Systems | Q1 – Q10
Section 2 | Ingesting and Processing Data | Q11 – Q20
Section 3 | Storing the Data | Q21 – Q30
Section 4 | Preparing and Using Data for Analysis | Q31 – Q40
Section 5 | Maintaining and Automating Data Workloads | Q41 – Q50
Instructions: Choose the single best answer. Pass: 75% in 90 minutes.
══════════════════════════════════════
SECTION 1: DESIGNING DATA PROCESSING SYSTEMS Q1 –
Q10
══════════════════════════════════════
Question 1 of 50
,2
A healthcare organization is migrating its on-premises data processing
system to Google Cloud to analyze sensitive patient records. The
architecture must ensure that Protected Health Information (PHI) is
encrypted at rest and in transit using keys managed entirely by the
healthcare provider, while also allowing the data to be processed by a
streaming pipeline.
A. Use Dataflow to process the streaming data, and configure Cloud
KMS Customer-Managed Encryption Keys (CMEK) for the underlying
Compute Engine disks and Pub/Sub topics. ✓ CORRECT
B. Use Dataproc to process the streaming data, and rely on Google-
managed encryption keys since Dataproc does not support custom
encryption keys for temporary storage.
C. Use the Transfer Appliance to physically transport the patient records
to Google Cloud, and apply application-layer encryption before writing
the data to BigQuery.
D. Use Cloud Run to ingest the data and write it directly to BigQuery,
and use VPC Service Controls to restrict access to the patient dataset.
Correct Answer: A
Rationale: Dataflow integrates natively with Cloud KMS to apply
Customer-Managed Encryption Keys (CMEK) to persistent disks and
Pub/Sub topics, ensuring PHI remains encrypted with customer-
controlled keys during streaming processing. Using Dataproc is incorrect
because while it does support CMEK for persistent disks, it is inherently
a batch processing system and requires extra complexity to handle
continuous streaming data effectively.
,3
Question 2 of 50
A global e-commerce company needs to maintain real-time inventory
counts across multiple warehouses worldwide. The system must provide
strongly consistent reads and writes to prevent overselling of products,
while sustaining thousands of concurrent transactions per second.
A. Use Cloud Bigtable with a multi-region cluster configuration to
ensure low-latency reads and writes for the inventory counters.
B. Use Cloud Spanner with a multi-region configuration to provide
strong consistency and horizontal scalability for transactional updates. ✓
CORRECT
C. Use BigQuery with frequently updated materialized views to
calculate the exact inventory counts across all global regions.
D. Use Cloud SQL for PostgreSQL with read replicas distributed across
the globe to handle the transactional workload.
Correct Answer: B
Rationale: Cloud Spanner is the only Google Cloud database that
combines strong global consistency with horizontal scalability, making it
ideal for high-concurrency, globally distributed transactional workloads
like inventory management. Using Cloud Bigtable is incorrect because it
only offers eventual consistency by default, which could lead to race
conditions and overselling in an inventory system.
Question 3 of 50
, 4
Your marketing team is storing five years of historical clickstream data
in BigQuery to generate annual performance reports. The data is
accessed only a few times a year, and the team wants to significantly
reduce storage costs while keeping the data queryable.
A. Convert the BigQuery tables to a compressed Avro format and move
them to Cloud Storage Standard class for direct querying via BigQuery
external tables.
B. Enable BigQuery long-term storage pricing on the tables by leaving
them untouched for 90 days, which automatically reduces the storage
cost by 50%.
C. Export the historical data to Cloud Storage Coldline, delete the
BigQuery tables, and create external tables pointing to the Coldline
buckets for the annual queries. ✓ CORRECT
D. Partition the BigQuery tables by year and cluster them by user ID to
reduce the amount of data scanned during the annual report generation.
Correct Answer: C
Rationale: Exporting infrequently accessed data to Cloud Storage
Coldline drastically cuts storage costs compared to BigQuery, while
BigQuery external tables allow you to query this archived data directly
without re-ingesting it. Relying solely on BigQuery long-term storage is
incorrect because it only provides a 50% discount, whereas moving truly
cold data to Coldline storage is significantly more cost-effective for data
accessed less than once a year.
Question 4 of 50
GCP Professional Data Engineer
Certification Exam Actual Exam
2026/2027 with Detailed Rationales |
Complete Exam-Style Questions | Pass
Guaranteed – A+ Graded
TABLE OF CONTENTS
Section 1 | Designing Data Processing Systems | Q1 – Q10
Section 2 | Ingesting and Processing Data | Q11 – Q20
Section 3 | Storing the Data | Q21 – Q30
Section 4 | Preparing and Using Data for Analysis | Q31 – Q40
Section 5 | Maintaining and Automating Data Workloads | Q41 – Q50
Instructions: Choose the single best answer. Pass: 75% in 90 minutes.
══════════════════════════════════════
SECTION 1: DESIGNING DATA PROCESSING SYSTEMS Q1 –
Q10
══════════════════════════════════════
Question 1 of 50
,2
A healthcare organization is migrating its on-premises data processing
system to Google Cloud to analyze sensitive patient records. The
architecture must ensure that Protected Health Information (PHI) is
encrypted at rest and in transit using keys managed entirely by the
healthcare provider, while also allowing the data to be processed by a
streaming pipeline.
A. Use Dataflow to process the streaming data, and configure Cloud
KMS Customer-Managed Encryption Keys (CMEK) for the underlying
Compute Engine disks and Pub/Sub topics. ✓ CORRECT
B. Use Dataproc to process the streaming data, and rely on Google-
managed encryption keys since Dataproc does not support custom
encryption keys for temporary storage.
C. Use the Transfer Appliance to physically transport the patient records
to Google Cloud, and apply application-layer encryption before writing
the data to BigQuery.
D. Use Cloud Run to ingest the data and write it directly to BigQuery,
and use VPC Service Controls to restrict access to the patient dataset.
Correct Answer: A
Rationale: Dataflow integrates natively with Cloud KMS to apply
Customer-Managed Encryption Keys (CMEK) to persistent disks and
Pub/Sub topics, ensuring PHI remains encrypted with customer-
controlled keys during streaming processing. Using Dataproc is incorrect
because while it does support CMEK for persistent disks, it is inherently
a batch processing system and requires extra complexity to handle
continuous streaming data effectively.
,3
Question 2 of 50
A global e-commerce company needs to maintain real-time inventory
counts across multiple warehouses worldwide. The system must provide
strongly consistent reads and writes to prevent overselling of products,
while sustaining thousands of concurrent transactions per second.
A. Use Cloud Bigtable with a multi-region cluster configuration to
ensure low-latency reads and writes for the inventory counters.
B. Use Cloud Spanner with a multi-region configuration to provide
strong consistency and horizontal scalability for transactional updates. ✓
CORRECT
C. Use BigQuery with frequently updated materialized views to
calculate the exact inventory counts across all global regions.
D. Use Cloud SQL for PostgreSQL with read replicas distributed across
the globe to handle the transactional workload.
Correct Answer: B
Rationale: Cloud Spanner is the only Google Cloud database that
combines strong global consistency with horizontal scalability, making it
ideal for high-concurrency, globally distributed transactional workloads
like inventory management. Using Cloud Bigtable is incorrect because it
only offers eventual consistency by default, which could lead to race
conditions and overselling in an inventory system.
Question 3 of 50
, 4
Your marketing team is storing five years of historical clickstream data
in BigQuery to generate annual performance reports. The data is
accessed only a few times a year, and the team wants to significantly
reduce storage costs while keeping the data queryable.
A. Convert the BigQuery tables to a compressed Avro format and move
them to Cloud Storage Standard class for direct querying via BigQuery
external tables.
B. Enable BigQuery long-term storage pricing on the tables by leaving
them untouched for 90 days, which automatically reduces the storage
cost by 50%.
C. Export the historical data to Cloud Storage Coldline, delete the
BigQuery tables, and create external tables pointing to the Coldline
buckets for the annual queries. ✓ CORRECT
D. Partition the BigQuery tables by year and cluster them by user ID to
reduce the amount of data scanned during the annual report generation.
Correct Answer: C
Rationale: Exporting infrequently accessed data to Cloud Storage
Coldline drastically cuts storage costs compared to BigQuery, while
BigQuery external tables allow you to query this archived data directly
without re-ingesting it. Relying solely on BigQuery long-term storage is
incorrect because it only provides a 50% discount, whereas moving truly
cold data to Coldline storage is significantly more cost-effective for data
accessed less than once a year.
Question 4 of 50
