Google Cloud · Free Practice Questions · Last reviewed May 2026
30real exam-style questions organised by domain, each with the correct answer highlighted and a plain-English explanation of why it's right — and why the others are wrong.
22% of exam · 6 sample questions below
A data engineer needs to design a stream processing pipeline that reads events from Pub/Sub, enriches them with data from a Cloud Storage file, and writes aggregated results to BigQuery. The pipeline must handle late-arriving events up to 1 hour. Which Dataflow feature should be used to manage late data?
Triggers
Watermarks
Watermarks track the event time progress and allowed lateness; Dataflow drops elements beyond the watermark.
Side inputs
Windowing
A company uses Dataproc to run daily Spark ML jobs. The jobs run for 2 hours each day. The team wants to reduce costs without changing job characteristics. Which strategy is MOST cost-effective?
Use a single-node cluster to eliminate overhead
Enable high-availability mode to avoid restarts
Use preemptible instances for worker nodes
Preemptible instances are cheap and Spark handles preemptions via fault tolerance.
Increase the number of standard workers to finish faster
A financial services company stream trades into Pub/Sub and processes with Dataflow. The pipeline must ensure exactly-once processing of each trade for regulatory compliance. However, Pub/Sub guarantees at-least-once delivery. Which combination of features should the Dataflow pipeline use to achieve exactly-once semantics?
Use Dataflow's exactly-once processing mode and implement idempotent writes in the sink
Dataflow's exactly-once mode with idempotent sinks ensures output exactly once.
Enable Pub/Sub message deduplication and use at-most-once delivery
Use global windowing and discard late data
Use Pub/Sub Lite with exactly-once delivery guarantee
A data engineer needs to create a BigQuery table that is partitioned by ingestion time and clustered by customer_id and transaction_date. They also want to limit access so that only users from a specific domain can query the table. Which approach should they use?
Create the table with partitioning only, then use a materialized view to restrict access
Create the table without clustering, use row-level security to filter by domain, and grant access to the table
Create the table with partitioning and clustering, then create an authorized view on the table and grant the view access to the domain users
Authorized views allow controlled access without granting direct table access.
Create the table with partitioning and clustering, then grant bigquery.dataViewer to the domain via IAM at the dataset level
A startup needs a fully managed, serverless Spark service to run occasional data processing jobs without managing clusters. They want to pay only for the resources used during job execution. Which Google Cloud service should they use?
Dataproc Serverless
Dataproc Serverless automatically manages resources for Spark jobs and charges per job.
Dataflow
Cloud Data Fusion
Dataproc
A company wants to use Cloud Data Fusion to build ETL pipelines. They need to connect to a legacy on-premises database using JDBC and also want to use prebuilt transforms from the Hub. Which two features should they use?
Cloud SQL JDBC driver and Cloud Functions
Dataproc Metastore and Cloud Storage sink
Wrangler and Dataproc
CDAP JDBC plugin and the Hub
CDAP JDBC plugin connects to on-prem DB; Hub provides prebuilt transforms.
Want more Designing Data Processing Systems practice?
Practice this domain25% of exam · 6 sample questions below
A data engineer needs to load 10 TB of CSV files from Amazon S3 into Google BigQuery on a daily basis. Which service should they use to automate this transfer?
Dataproc
Cloud Data Fusion
BigQuery Data Transfer Service
BigQuery Data Transfer Service supports scheduled transfers from Amazon S3 directly into BigQuery.
Storage Transfer Service
You need to stream real-time user click events from your application into BigQuery for immediate analysis. The events must be available for query within seconds. Which approach is recommended?
Use Pub/Sub to Dataflow to BigQuery with the Storage Write API for high-throughput streaming.
This is the recommended architecture: Pub/Sub for ingestion, Dataflow for stream processing, and Storage Write API for low-latency streaming writes.
Use Cloud Data Fusion to ingest streaming data from Pub/Sub into BigQuery.
Use Cloud Functions to receive events from Pub/Sub and insert them into BigQuery using the legacy streaming API.
Use Pub/Sub with a BigQuery subscription to directly write events into BigQuery.
A data engineer needs to transfer 500 TB of on-premises data to Google Cloud Storage. The data is stored on NAS devices and the network bandwidth is limited to 100 Mbps. What is the most cost-effective and timely transfer method?
Use Storage Transfer Service over the internet
Use a VPN connection and rsync
Use gsutil cp in parallel
Use Transfer Appliance
Transfer Appliance is designed for offline petabyte-scale transfers, avoiding bandwidth limitations.
You are building a Dataflow pipeline in Python that reads messages from Pub/Sub, enriches them with data from a BigQuery table, and writes the results to BigQuery. The enrichment lookup table is large and changes infrequently. Which approach minimizes cost and latency?
Use a CoGroupByKey transform to join the incoming stream with a stream from BigQuery.
Use BigQuery IO to query the table for every incoming message.
Use a side input that reads the BigQuery table periodically and caches it.
Side inputs are ideal for distributing a static lookup table to all workers. The data can be refreshed on a schedule.
Use a stateful DoFn and store the lookup in state per key.
You are designing a Dataflow pipeline to process streaming data. The pipeline may encounter malformed records. You need to handle these errors without failing the entire pipeline and store the bad records for later analysis. What is the best practice?
Use a dead letter sink to write malformed records to a separate Pub/Sub topic or GCS location.
This is the recommended pattern: isolate bad records for later reprocessing while allowing the pipeline to continue.
Catch the exception and log it, then continue processing.
Write all records to BigQuery using the Storage Write API and handle errors in the write operation.
Raise an exception in the DoFn to stop the pipeline for manual intervention.
Your company uses Kafka for event streaming. You want to run Kafka on Google Cloud with the ability to auto-scale clusters and use managed infrastructure. Which service should you choose?
Cloud Pub/Sub
Confluent Cloud on GCP
Cloud Dataflow
Dataproc
Dataproc supports running Kafka as an optional component on managed clusters, giving you control and scalability.
Want more Ingesting and Processing the Data practice?
Practice this domainA company needs a fully managed, globally distributed relational database with strong consistency, external consistency, and 99.999% SLA for a financial transaction processing system. Which Google Cloud service should they use?
Firestore
Cloud Spanner
Cloud Spanner is globally distributed, strongly consistent, and offers 99.999% SLA.
Bigtable
Cloud SQL
A data engineer needs to store raw sensor data in Cloud Storage and automatically transition it to a lower-cost storage class after 30 days, then delete it after 365 days. What should they configure?
Use Cloud Pub/Sub notifications to trigger a Cloud Function that moves objects.
Use gsutil rewrite command in a cron job.
Configure a lifecycle rule with SetStorageClass to Nearline after 30 days and Delete after 365 days.
Lifecycle rules can automatically transition objects to a different storage class and then delete them based on age.
Set a bucket retention policy with a retention period of 365 days.
A data engineer is building a data lake on Google Cloud and needs to separate raw ingested data, curated/cleaned data, and processed/aggregated data. Which Cloud Storage bucket structure is recommended?
Create three separate folders in a single bucket: raw, curated, processed.
Using prefixes (folders) within a bucket is a standard pattern for organizing data lake zones, allowing different lifecycle rules per prefix.
Store all data in one bucket and use object labels to distinguish raw, curated, and processed.
Store raw data in a different project for security isolation.
Use different storage classes for raw, curated, and processed data within the same bucket.
A mobile app needs an offline-first NoSQL database that syncs data across devices when connectivity is available. Which Google Cloud database meets these requirements?
Memorystore
Cloud SQL
Bigtable
Firestore
Firestore provides offline data persistence and automatic sync, perfect for mobile apps.
An organization needs to prevent data exfiltration from BigQuery by ensuring all traffic to BigQuery APIs goes through VPC boundaries and is restricted to a specific service perimeter. Which Google Cloud security control should they use?
Access Transparency
IAM conditions on BigQuery roles
Cloud Armor
VPC Service Controls
VPC Service Controls define a perimeter that restricts data movement to authorized networks and prevents exfiltration.
A data engineer needs to design a Bigtable row key for a time-series IoT application where each device sends data every second. The query pattern is to retrieve all data for a specific device over a time range. Which row key design minimizes hotspots?
device_id#timestamp (e.g., device123#2024-03-15-10:30:00)
hash(device_id)#timestamp (e.g., a3f2#2024-03-15-10:30:00)
Hashing the device ID distributes writes across tablets, and appending timestamp allows efficient time-range scans.
timestamp#device_id (e.g., 2024-03-15-10:30:00#device123)
device_type#device_id#timestamp
Want more Storing the Data practice?
Practice this domain15% of exam · 6 sample questions below
A data engineer wants to train a linear regression model in BigQuery ML to predict sales. The training data includes a categorical feature with 1000+ unique values. Which method is most appropriate to handle this feature in the CREATE MODEL statement?
Set max_categorical_features=100 in the model options.
Use TRANSFORM clause with ML.FEATURE_CROSS or manual hashing.
TRANSFORM allows custom feature engineering including hashing for high-cardinality features.
Use the OPTIONS(ENCODE='ONE_HOT_ENCODING') parameter in the model options.
The model automatically handles high-cardinality features without any additional steps.
You need to create a Looker model that defines a 'sales' view based on a BigQuery table, with a measure for total revenue. Which LookML object defines the table and dimensions?
explore
view
A view in LookML maps to a database table and defines dimensions and measures.
model
dimension
A company uses Looker Studio to build dashboards from BigQuery data. They notice that queries take several seconds to return. They want to improve performance without changing the schema or adding materialized views. Which option should they use?
Enable BigQuery BI Engine on the relevant project.
BI Engine provides in-memory analysis for Looker Studio, reducing query latency.
Move the data to Cloud SQL.
Switch to BigQuery Omni for cross-cloud queries.
Use APPROX_COUNT_DISTINCT to speed up distinct counts.
A data scientist is training a binary classification model on an imbalanced dataset (95% negative, 5% positive) using AutoML Tables. Which strategy should they use to handle the class imbalance?
Set the budget to a higher value to allow more training on minority class.
Use SMOTE in a Dataflow pipeline before importing the data to AutoML Tables.
Specify a weight column with higher weights for positive examples in the dataset.
AutoML Tables supports a weight column to give more importance to minority class.
Create duplicate copies of the positive class rows to balance the dataset.
You need to split a time-series dataset into training and evaluation sets for a forecasting model. The data is ordered by timestamp. Which splitting technique should you use?
Sequential split where training data precedes evaluation data in time.
Sequential split respects the temporal order and prevents leakage.
Use k-fold cross-validation with random folds.
Stratified split based on the target variable.
Random split with 80% training, 20% evaluation.
Which BigQuery SQL function returns the rank of a row within a window, with gaps in the ranking for ties?
RANK()
RANK() handles ties with gaps.
NTILE()
DENSE_RANK()
ROW_NUMBER()
Want more Preparing and Using Data for Analysis practice?
Practice this domain18% of exam · 6 sample questions below
A data engineer uses Cloud Composer to orchestrate a daily batch pipeline. A downstream task should only start after an upstream BigQuery load job finishes successfully and a specific file appears in Cloud Storage. Which combination of operators should the engineer use in the Airflow DAG?
BigQueryInsertJobOperator with wait_for_downstream=True
BigQueryInsertJobOperator and GCSObjectExistenceSensor with upstream dependency
Correct: BigQueryInsertJobOperator performs the load, GCSObjectExistenceSensor polls for the file, and upstream dependency ensures order.
DataflowPythonOperator and GCSObjectExistenceSensor
BigQueryOperator and FileSensor with downstream dependency
A company uses Dataflow streaming pipelines to process real-time events. They notice increasing system lag over time. Which two Cloud Monitoring metrics should be examined to diagnose the cause?
Pub/Sub subscription/num_undelivered_messages and Dataflow job/watermark_lag
Dataproc cluster/yarn_allocated_memory_percentage and Dataflow job/worker_cpu
Dataflow job/system_lag and Dataflow job/data_freshness
system_lag indicates processing delay; data_freshness shows watermark progress. Both are key for streaming lag.
BigQuery query/execution_times and Dataflow job/elapsed_time
A data team needs to share a BigQuery dataset with another business unit. They want to provide a point-in-time snapshot of the data without incurring additional storage costs for the copy. Which BigQuery feature should they use?
BigQuery table snapshots
BigQuery table clones
Clones are writable and share storage with the base table, so no extra cost for the initial copy. They can be updated independently.
BigQuery authorized views
BigQuery export to Cloud Storage
An engineer needs to create a reusable Dataflow pipeline that can be executed with different parameters without modifying code. Which Dataflow feature should they use?
Dataflow Shuffle
Dataflow Flex Templates
Flex Templates use Docker containers and support any pipeline dependency, allowing parameterization.
Dataflow SQL
Dataflow Classic Templates
A company runs a Dataproc cluster for ETL jobs that process data nightly. They want to reduce costs while maintaining performance. Which strategy is MOST effective?
Use committed use discounts for all VMs
Enable Dataproc auto-scaling
Use preemptible VMs for all nodes including master
Use preemptible VMs for worker nodes only
Workers can be preemptible because batch jobs can tolerate interruptions; master remains on-demand for reliability.
A data engineer needs to alert when Pub/Sub subscription has messages older than 1 hour. Which Cloud Monitoring metric and filter should they use?
Metric: topic/send_message_operation_count; filter: topic_id
Metric: subscription/ack_message_count; filter: subscription_id
Metric: subscription/num_undelivered_messages; filter: subscription_id
Metric: subscription/oldest_unacked_message_age; filter: subscription_id
Correct metric and filter for alerting on message age.
Want more Maintaining and Automating Data Workloads practice?
Practice this domainThe PDE exam has 60 questions and must be completed in 120 minutes. The passing score is 720/1000.
Scenario-based questions covering exam objectives with detailed answer explanations.
The exam covers 5 domains: Designing Data Processing Systems, Ingesting and Processing the Data, Storing the Data, Preparing and Using Data for Analysis, Maintaining and Automating Data Workloads. Questions are weighted by domain — higher-weight domains appear more on your actual exam.
No. These are original exam-style practice questions written against the official Google Cloud PDE exam objectives. They are not copied from the real exam. Courseiva focuses on genuine understanding, not memorisation of braindumps.
Courseiva tracks your accuracy per domain and routes you toward weak areas automatically. Free, no account required.