PDE Practice Test 36 — 15 Questions

Question 1

A team needs to run hybrid transactional/analytical workloads on PostgreSQL-compatible data with low latency. They require high performance on both OLTP and OLAP queries, leveraging a columnar engine. Which Google Cloud service is best suited?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

AlloyDB. AlloyDB is the correct choice because it is a fully managed PostgreSQL-compatible database service on Google Cloud that combines a columnar engine for fast analytical queries with high transactional performance. It uses a columnar query accelerator to offload analytical workloads from the transactional engine, enabling low-latency hybrid transactional/analytical processing (HTAP) without data movement.

Answer

Cloud SQL for PostgreSQL

Answer

Cloud Spanner

Answer

BigQuery

Question 2

A company needs to store and analyze large amounts of unstructured data (images, videos) and structured data (CSV logs) in a cost-effective manner. The data should be accessible for analytics with BigQuery. Which two services should they use? (Choose TWO.)

Accepted Answer

BigQuery. Cloud Storage is the best option for storing unstructured and structured files cost-effectively. BigQuery can analyze this data directly via external tables or after loading, making it a powerful analytics platform.

Answer

Cloud SQL

Answer

Cloud Spanner

Answer

Firestore

Question 3

A company uses BigQuery for analytics. They need to ensure data quality by preventing duplicate records from being inserted. Which approach is most effective?

Accepted Answer

Use a DML MERGE statement that filters out duplicates based on a unique key.. BigQuery's DML MERGE statement can be used to atomically insert rows only when a unique key does not already exist in the target table. By using a MERGE with a WHEN NOT MATCHED THEN INSERT clause, the operation prevents duplicate records from being inserted in a single, transactional statement, ensuring data quality without requiring external tools or post-processing.

Answer

Use BigQuery ML to train a model that identifies anomalies.

Answer

Use Cloud Data Loss Prevention API to scan for duplicates.

Answer

Use COUNT DISTINCT in queries to ignore duplicates.

Question 4

Which THREE metrics should be monitored to detect model drift in a production ML system?

Accepted Answer

Prediction distribution (prediction drift).. Option B (prediction distribution / prediction drift) is correct because monitoring how the model's output scores or predicted labels shift over time reveals changes in model behavior even before ground truth labels arrive, which is a core signal of model drift. Option C (feature distribution / data drift) is correct because drift in input features relative to the training distribution (e.g., via PSI, KL divergence, or KS tests) is a leading indicator that the model is being fed data unlike what it learned from. Option E (model performance metrics such as accuracy, precision, recall on a ground truth dataset) is correct because once labeled outcomes are available, degradation in these metrics directly confirms that the model's predictive quality has decayed. Option A (training loss convergence) is not a drift metric—it is a one-time training diagnostic and does not reflect post-deployment data changes. Option D (CPU utilization of serving nodes) is an infrastructure/operational metric that indicates resource pressure, not statistical or behavioral drift in the model.

Answer

Training loss convergence.

Answer

CPU utilization of the serving nodes.

Question 5

Which TWO statements are true about BigQuery Data Transfer Service? (Choose 2)

Accepted Answer

It supports scheduled transfers from Amazon S3 and Redshift.. Option D is correct because BigQuery Data Transfer Service provides built-in connectors for Amazon S3 and Amazon Redshift, enabling recurring, scheduled (and on-demand) transfers of external data into BigQuery without custom code. Option E is correct because the service also offers first-party Google connectors, including Google Ads, YouTube (Channel/Content Owner reports), and Google Ad Manager, which load those marketing datasets into BigQuery on a schedule. Option A is not correct because the Data Transfer Service is designed for ingestion/movement, not in-flight transformation; transformations are typically done after loading using BigQuery SQL, views, or scheduled queries. Option B is not correct because Cloud SQL is not a native Data Transfer Service source; moving Cloud SQL data into BigQuery generally requires tools such as Datastream, the Cloud SQL federated query/EXTERNAL_QUERY, or export/load pipelines. Option C is not correct because BigQuery Data Transfer Service is available in many more locations than just US and EU, and its regional availability follows BigQuery dataset locations rather than being limited to those two regions.

Answer

It supports data transformation during transfer.

Answer

It can transfer data from Cloud SQL to BigQuery directly.

Answer

It is only available in the US and EU regions.

Question 6

A Dataflow pipeline reads from Cloud Pub/Sub and writes to Cloud Storage. The pipeline needs to guarantee exactly-once processing despite worker failures. Which configuration ensures exactly-once semantics?

Accepted Answer

Enable exactly-once delivery on the Pub/Sub subscription and use an idempotent sink. Pub/Sub subscriptions can be configured with exactly-once delivery (using the `enableExactlyOnceDelivery` flag), which ensures that each message is delivered to the subscriber exactly once. Combining this with an idempotent sink (e.g., Cloud Storage with unique filenames or deduplication logic) guarantees that even if a worker fails and the pipeline retries, the output will not contain duplicates. This is the only option that directly addresses both the source and sink to achieve end-to-end exactly-once semantics.

Answer

Use a side input from a deduplication dataset

Answer

Set the pipeline to use a global window with no early triggers

Answer

Insert a Reshuffle transform after reading

Question 7

After migrating a production Cloud SQL for PostgreSQL database to a larger machine type, the team notices slower queries. What is the best step to identify the cause?

Accepted Answer

Enable pg_stat_statements and review query execution times.. Pg_stat_statements is a PostgreSQL extension that provides detailed query execution statistics, including total execution time, number of calls, and I/O metrics. After migrating to a larger machine type, slower queries often stem from plan changes due to different hardware characteristics or configuration settings; reviewing pg_stat_statements output helps pinpoint which queries are underperforming and why.

Answer

Reindex all tables to improve index efficiency.

Answer

Enable query caching through the database flags.

Answer

Increase max_connections to handle more concurrent queries.

Question 8

A healthcare startup is deploying a natural language processing (NLP) model for extracting medical entities from clinical notes. The model is a fine-tuned BERT model served on Vertex AI Prediction using a custom container. The team observes that prediction latency is around 500ms per request, but they need to handle up to 100 requests per second (QPS) with end-to-end latency under 200ms. The model currently runs on n1-standard-4 machines (4 vCPU, 15 GB memory). During load testing, CPU utilization reaches 90% and memory usage is 12 GB. The team is considering options to meet the requirements. Which action should they take?

Accepted Answer

Use a machine type with a GPU, such as n1-standard-4 with a NVIDIA Tesla T4 accelerator, and optimize the model with TensorRT.. The bottleneck is CPU-bound inference (90% CPU utilization) with memory well within limits (12 GB of 15 GB). Adding a GPU (NVIDIA Tesla T4) and optimizing with TensorRT reduces per-request latency via hardware acceleration and graph optimizations, enabling sub-200ms inference at 100 QPS. This directly addresses the latency requirement without changing the machine family or scaling strategy.

Answer

Switch to n1-highmem-4 machines to provide more memory for the model.

Answer

Deploy the model using TensorFlow Serving with CPU-only nodes and increase the number of replicas.

Answer

Move the model to Cloud Run with automatic scaling to handle the QPS.

Question 9

A company wants to use dbt to transform data in BigQuery. Their source data is loaded daily into staging tables. They need to run dbt transformations on a schedule and only process tables that have changed. Which dbt feature should they use?

Accepted Answer

dbt incremental models. dbt incremental models are designed to process only new or changed data since the last run, rather than rebuilding the entire table. This is ideal for daily transformations on staging tables where only some data changes. Incremental models use a unique key and a strategy (e.g., merge, delete+insert) to update the target table efficiently.

Answer

dbt snapshots

Answer

dbt seeds

Answer

dbt sources

Question 10

Which TWO features of Cloud Pub/Sub guarantee at-least-once delivery and enable exactly-once processing downstream? (Choose two.)

Accepted Answer

Subscriber-retry policy with exponential backoff.. Option A is correct because a subscriber-retry policy with exponential backoff ensures that when a message handler fails or does not acknowledge in time, Pub/Sub redelivers the message with increasing delays, which is a core mechanism supporting at-least-once delivery. Option E is correct because acknowledgment deadlines determine how long Pub/Sub waits for an ack before redelivering, and message persistence (retaining messages in storage until acknowledged or expired) guarantees that messages are not lost, thereby enforcing at-least-once delivery. Together, these features let downstream consumers deduplicate or use idempotent processing to achieve exactly-once processing semantics. Option B is not marked correct because exactly-once delivery is an optional subscription setting, not enabled by default, and the question asks for features guaranteeing at-least-once delivery. Option C is not correct because message ordering by key only preserves order within an ordering key and does not itself guarantee delivery or exactly-once processing. Option D is not correct because Cloud Dataproc integration is unrelated to Pub/Sub delivery guarantees or replay semantics.

Answer

Exactly-once delivery subscription setting (optional, not enabled by default)

Answer

Message ordering by message key.

Answer

Cloud Dataproc integration for message replay.

Question 11

You are designing a data pipeline for ML training with Vertex AI. You need to split time-series data into train/validation/test sets without leaking future data. Which THREE practices should you follow?

Accepted Answer

Use a sliding window validation approach for hyperparameter tuning.. Option A is correct because a sliding-window (rolling-origin) validation scheme trains on past data and validates on the immediately following window, which respects temporal order and prevents future leakage during hyperparameter tuning. Option B is correct because keeping all data points from the same time period (e.g., same day or timestamp bucket) in a single split avoids boundary leakage, where records from one period appear in both train and test. Option E is correct because using an explicit date column to define split boundaries (for example, train on data before 2024-01-01, validate on Q1 2024, test on Q2 2024) enforces a strict chronological cutoff. Option C is not appropriate because Looker is a BI/analytics and dashboarding tool, not a mechanism for producing leakage-safe ML dataset splits. Option D is wrong because random row assignment mixes past and future observations across splits, which is exactly the temporal leakage the scenario must avoid.

Answer

Use Looker to generate the splits automatically.

Answer

Randomly assign rows to each split to ensure statistical distribution.

Question 12

To enable data lineage tracking in BigQuery, which feature should be activated?

Accepted Answer

Dataplex Lineage. The correct feature for data lineage tracking in BigQuery is Dataplex Lineage. Dataplex provides end-to-end lineage tracking across data assets, including BigQuery tables and views. BigQuery Lineage API is not a distinct feature; lineage capabilities are integrated into Dataplex. BigQuery Audit Logs capture metadata changes but are not lineage-specific, and Data Catalog is for metadata management, not lineage.

Answer

BigQuery Audit Logs

Answer

Data Catalog

Answer

BigQuery Lineage API

Question 13

A team is using Kubeflow Pipelines on Google Kubernetes Engine to orchestrate ML workflows. They need to track parameters, metrics, and artifacts for each run. Which tool should they integrate?

Accepted Answer

Vertex ML Metadata. Vertex ML Metadata is the correct choice because it is purpose-built for tracking parameters, metrics, and artifacts in ML workflows, and it integrates natively with Kubeflow Pipelines on Google Kubernetes Engine. It stores metadata for each pipeline run, enabling lineage tracking, comparison, and reproducibility of experiments.

Answer

Cloud Monitoring

Answer

Cloud Logging

Answer

BigQuery

Question 14

Your team runs a Cloud Composer 2 environment to orchestrate BigQuery ELT jobs. A DAG that loads a critical fact table must only run after an upstream ingestion DAG completes, and you want Composer itself to trigger it automatically without any external scheduler. Which mechanism should you use?

Accepted Answer

An ExternalTaskSensor that watches the upstream DAG's task in the same Composer environment.. Airflow's ExternalTaskSensor exists specifically to express a dependency on a task in another DAG within the same Airflow deployment. It checks the external DAG run for the matching execution date and only allows downstream tasks to proceed after the upstream task succeeds. Time-based waits, external cron triggers, and message-based glue all fail to guarantee that the ingestion actually finished successfully before the load begins.

Answer

A TimeDeltaSensor with a fixed delay that matches the upstream DAG's typical runtime.

Answer

A Pub/Sub push subscription that invokes a Cloud Function to call the downstream DAG's trigger endpoint.

Answer

A Cloud Scheduler job that calls the Airflow REST API to trigger the downstream DAG on a cron schedule.

Question 15

You are building a binary classification model using AutoML Tables on Vertex AI. The dataset has a severe class imbalance (1% positive class). Which strategy should you use to handle the imbalance?

Accepted Answer

Use a weight column to assign higher weights to the minority class.. The correct option is D: use a weight column in AutoML Tables to assign higher weights to the minority class. In Vertex AI AutoML Tables, you can add a weight column to the training data and specify it as the weight column, so minority-class rows contribute more to the loss. This directly counteracts the 1% positive-class imbalance without changing the data distribution. Options A and C are not supported preprocessing steps in AutoML Tables, since it manages data splitting and training internally and does not expose SMOTE or manual downsampling. Option B is incorrect because AutoML Tables does not automatically correct severe class imbalance; you must supply weights or adjust the optimization objective. Option D is therefore the appropriate, supported mechanism for this scenario.

Answer

Oversample the minority class using SMOTE before training.

Answer

Do nothing; AutoML Tables automatically handles class imbalance.

Answer

Downsample the majority class to match the minority class size.