PDE Practice Test 23 — 15 Questions

Question 1

You want to train a custom TensorFlow model on Vertex AI using a managed Jupyter notebook environment. Which service should you use?

Accepted Answer

Vertex AI Workbench. Vertex AI Workbench provides managed Jupyter notebooks with pre-installed frameworks and easy access to Vertex AI services.

Answer

Cloud Datalab

Answer

Vertex AI Training

Answer

AI Platform Notebooks

Question 2

A data engineer is designing a BigQuery table for a clickstream dataset with frequent queries aggregating over user sessions. Each user session has multiple events, and the engineer wants to avoid joins for performance. Which schema design pattern should they use?

Accepted Answer

Use nested and repeated fields to store events within each session row. BigQuery supports nested and repeated fields (e.g., STRUCT and REPEATED), allowing denormalization. This reduces the need for joins and improves query performance. Partitioning and clustering are for physical data organization, not schema design.

Answer

Use a normalized schema with separate tables for sessions and events, then join on session ID

Answer

Store each event as a separate row with session key and use clustering on session ID

Answer

Use partitioning on event timestamp and clustering on user ID

Question 3

A retail company wants to trigger a Cloud Run service whenever a new CSV file is uploaded to a specific Cloud Storage bucket. Which THREE components are needed to set up this event-driven architecture? (Choose 3)

Accepted Answer

Cloud Storage bucket with notifications enabled. Cloud Storage buckets must have notifications enabled to publish events to Pub/Sub when objects are created. Without enabling notifications, the bucket cannot emit events that trigger downstream services. This is typically done by configuring the bucket to send notifications to a Pub/Sub topic for each new object upload.

Answer

Cloud Dataflow pipeline

Answer

Cloud Pub/Sub topic

Question 4

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?

Accepted Answer

view. In LookML, a view defines the mapping to a database table (or derived table) and contains dimensions and measures.

Answer

explore

Answer

model

Answer

dimension

Question 5

A data engineer is building a production ML pipeline on Vertex AI. The pipeline must preprocess features (e.g., scaling, encoding) and then train a model. The preprocessing logic must be reusable for serving predictions. Which Vertex AI component should they use?

Accepted Answer

Vertex AI Feature Transform. Vertex AI Feature Transform is a managed service that allows you to define transformations using TFX Transform or BigQuery SQL, which are then applied consistently during training and serving. Vertex AI Pipelines can orchestrate but does not itself provide reusable transformations. Dataflow would require custom code. Vertex AI Feature Store serves pre-computed features, not transformations.

Answer

Dataflow with Apache Beam

Answer

Vertex AI Feature Store

Answer

Vertex AI Pipelines

Question 6

You need to track the lineage of data in BigQuery, showing how tables are derived from other tables via queries. Which service provides this capability?

Accepted Answer

BigQuery Lineage API. BigQuery lineage API and Dataplex lineage both provide data lineage tracking. BigQuery lineage is built-in, while Dataplex extends it across the data lake.

Answer

Cloud Composer

Answer

Cloud Data Catalog

Answer

Dataflow

Question 7

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?

Accepted Answer

Metric: subscription/oldest_unacked_message_age; filter: subscription_id. The metric subscription/oldest_unacked_message_age gives the age of the oldest unacknowledged message. Filtering by subscription ID targets the specific subscription.

Answer

Metric: topic/send_message_operation_count; filter: topic_id

Answer

Metric: subscription/ack_message_count; filter: subscription_id

Answer

Metric: subscription/num_undelivered_messages; filter: subscription_id

Question 8

An IoT application writes sensor readings to Cloud Bigtable with a row key of 'deviceID#timestamp'. The team notices high write latency and hotspots on a few nodes. Which row key design change would most likely improve performance?

Accepted Answer

Add a random prefix to the row key (e.g., hash of deviceID modulo 1000). Prefixing timestamps can cause hotspots because writes go to the same tablet server for the same time range. Hashing the device ID or using a field-leveled design (e.g., deviceID inverted) distributes writes across nodes. Adding a random prefix helps but salting with a hash is more systematic.

Answer

Reverse the key to 'timestamp#deviceID'

Answer

Use a single column family with many columns

Answer

Store all data in one row per device

Question 9

A data engineer wants to automatically move objects from Standard storage class to Nearline after 30 days, and then to Archive after 365 days. Which Cloud Storage feature should they configure?

Accepted Answer

Object Lifecycle rule with SetStorageClass actions. Object Lifecycle rules in Google Cloud Storage allow you to automatically transition objects between storage classes (e.g., from Standard to Nearline after 30 days, then to Archive after 365 days) using the SetStorageClass action. This feature is specifically designed for automated lifecycle management, including deletion and class transitions, based on object age or other conditions.

Answer

Object Versioning

Answer

Retention Policy

Answer

Bucket Lock

Question 10

A financial analytics team uses Looker to explore BigQuery data. They need to allow business users to filter by a custom date range that is not tied to an existing dimension. The date range must be user-input at query time. What is the best approach in Looker?

Accepted Answer

Create a parameter in LookML using Liquid templating. Looker uses Liquid templating in LookML to create parameters that render as filter controls at runtime. Users can input values that are then injected into the SQL. Creating a dimension with a yesno filter requires predefined values. The filter parameter on a dimension only allows selecting from existing values, not arbitrary input.

Answer

Create an explore with a custom filter field in the Looker UI

Answer

Use a filter parameter directly on the date dimension

Answer

Add a dimension with a yesno filter that toggles the date range

Question 11

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?

Accepted Answer

Specify a weight column with higher weights for positive examples in the dataset.. AutoML Tables automatically handles class imbalance by applying class weights and downsampling. Users can also specify a weight column explicitly.

Answer

Set the budget to a higher value to allow more training on minority class.

Answer

Use SMOTE in a Dataflow pipeline before importing the data to AutoML Tables.

Answer

Create duplicate copies of the positive class rows to balance the dataset.

Question 12

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?

Accepted Answer

Create three separate folders in a single bucket: raw, curated, processed.. A common best practice for data lakes on GCS is to use separate buckets or folders within a bucket (e.g., raw, curated, processed) to manage different stages of data refinement and apply appropriate lifecycle policies.

Answer

Store all data in one bucket and use object labels to distinguish raw, curated, and processed.

Answer

Store raw data in a different project for security isolation.

Answer

Use different storage classes for raw, curated, and processed data within the same bucket.

Question 13

You are optimizing a BigQuery query that scans 1 TB of data every day. The query joins a large fact table (partitioned by date) with a small dimension table. You notice that the query always scans the entire fact table, even though you only need the last 7 days of data. Which optimization will MOST reduce the bytes scanned?

Accepted Answer

Add a WHERE clause that filters on the date column used for partitioning.. Using a WHERE clause on the partitioning column (e.g., date) allows BigQuery to perform partition pruning, drastically reducing scanned bytes. Clustering on the join key can improve performance but does not reduce bytes scanned as much as partition pruning. Materialized views precompute aggregations but may not reduce scans if the base query still scans full table. Changing to a clustered table does not prune partitions automatically.

Answer

Create a materialized view that pre-aggregates the data by day.

Answer

Cluster the fact table on the join key used in the query.

Answer

Change the table to use time-unit column partitioning with a 1-day partition interval.

Question 14

You are running a streaming pipeline with Dataflow that reads from Pub/Sub and writes to BigQuery. You notice that the system lag metric is increasing over time, indicating that messages are taking longer to process. What is the most likely cause and how should you address it?

Accepted Answer

The Dataflow workers are CPU-bound; increase the number of workers or adjust autoscaling settings.. Increasing system lag often means the pipeline is CPU-bound, failing to keep up with the incoming data rate. Updating the pipeline with a higher number of workers (or enabling autoscaling) can resolve this.

Answer

The source Pub/Sub topic has insufficient throughput; increase the number of partitions.

Answer

The BigQuery destination table has too many columns; reduce the number of columns.

Answer

The pipeline uses a batch transform that should be replaced with a streaming transform.

Question 15

A data engineer needs to query data across BigQuery (in Google Cloud) and Snowflake (in AWS) without moving the data. Which service should they use?

Accepted Answer

BigQuery Omni. BigQuery Omni allows querying data across multiple clouds using BigQuery's interface, with compute running in the respective cloud. Data stays in place.

Answer

Dataflow

Answer

Cloud SQL

Answer

Vertex AI Feature Store