GCP-ADP Practice Test 13 — 15 Questions

Question 1

Which TWO factors should you consider when choosing a GCS storage class?

Accepted Answer

Data retention requirements. Option B (Data retention requirements) is correct because GCS storage classes differ in minimum storage duration and retrieval costs, so how long you must keep data determines whether a class like Nearline, Coldline, or Archive is cost-effective versus Standard. Option E (Access frequency) is correct because storage classes are fundamentally optimized around how often objects are read: Standard for frequent access, Nearline for about once a month, Coldline for about once a quarter, and Archive for less than once a year. Option A (Data color) is not a real GCS concept and has no bearing on storage class selection. Option C (Network speed) affects transfer performance but not which storage class you choose. Option D (Number of users) is irrelevant because GCS storage classes are based on access patterns and retention, not user counts.

Answer

Data color

Answer

Network speed

Answer

Number of users

Question 2

You are using Airflow with a Cloud Composer environment. You want to ensure that a specific task runs on a specific day of the week. How do you configure the schedule?

Accepted Answer

Set 'schedule_interval' using cron syntax.. In Airflow (and thus Cloud Composer), the schedule of a DAG is controlled by the 'schedule_interval' parameter, which accepts cron expressions, timedelta objects, or presets like '@daily'. To run a task on a specific day of the week, you use a cron expression such as '0 0 * * 1' for every Monday. This is the standard, declarative way to define weekly schedules in Airflow.

Answer

Create a separate DAG for each day of the week.

Answer

Enable the 'weekday_only' flag in the DAG configuration.

Answer

Use the 'timedelta' object with a 7-day offset.

Question 3

You need to control access to specific columns in a BigQuery table using Dataplex-integrated policy tags. What must you create before you can apply a policy tag to a column?

Accepted Answer

Policy Tag Taxonomy. Policy tags in BigQuery are organized within a Policy Tag Taxonomy, which is a hierarchical container created in Data Catalog (now Dataplex Universal Catalog). Before any policy tag can be applied to a column, you must first create a taxonomy to hold the tags, then create the individual policy tags inside it, and finally grant the Fine-Grained Reader role to users. The taxonomy is the mandatory parent resource — you cannot create a standalone policy tag without it.

Answer

Data Zone

Answer

IAM Condition

Answer

BigQuery Dataset permission

Question 4

Which of the following is NOT a benefit of using Dataplex?

Accepted Answer

High-frequency OLTP transaction support. Dataplex is a data governance and metadata management service designed for analytics workloads — it provides automated discovery, cataloging, data quality, and unified management across data lakes. It is not an OLTP (online transaction processing) engine; it does not support high-frequency transactional workloads like row-level inserts/updates at scale. Therefore, high-frequency OLTP transaction support is NOT a benefit of Dataplex.

Answer

Automated data discovery

Answer

Improved data compliance

Answer

Unified data management

Question 5

You are building a Looker dashboard and need to calculate the year-over-year growth rate. Which feature should you use to perform this without modifying the underlying SQL?

Accepted Answer

Looker Table Calculations.. Looker Table Calculations allow you to create calculations (such as year-over-year growth) directly in the Explore or dashboard using Looker's expression language, without modifying the underlying SQL or LookML. They operate on the query results already returned, so you can compute percent change, offsets, and running totals on the fly. This makes them ideal for ad-hoc metrics like YoY growth.

Answer

BigQuery query parameters.

Answer

LookML derived tables with persistent caching.

Answer

Database-level User Defined Functions (UDFs).

Question 6

In BigQuery, you want to improve query performance for a table that is frequently filtered by a `customer_id` column. What is the recommended strategy?

Accepted Answer

Cluster the table by 'customer_id'.. Clustering a BigQuery table by customer_id physically sorts the data by that column, so queries with filters on customer_id can prune blocks and scan far less data. This is the recommended strategy for improving filter performance on a frequently queried column. Clustering is especially effective for high-cardinality columns and works well with partitioning.

Answer

Add a primary key constraint.

Answer

Convert the table to a View.

Answer

Increase the number of slots in the reservation.

Question 7

You need to perform a complex windowing operation on streaming data in Dataflow. Which windowing strategy is best for session-based activity?

Accepted Answer

Session windows. Session windows are designed for session-based activity, where data is grouped by periods of activity separated by gaps of inactivity. They dynamically merge windows based on a specified gap duration, making them ideal for analyzing user sessions. This matches the requirement for complex windowing on streaming data with session-based activity.

Answer

Fixed windows

Answer

Global windows

Answer

Sliding windows

Question 8

A user complains they cannot see a Dataplex Lake in the console. You verify they have the 'Viewer' role on the project. What is missing?

Accepted Answer

They need the 'roles/dataplex.viewer' role. Dataplex uses its own IAM roles, and project-level Viewer (roles/viewer) does not grant permission to view Dataplex lakes. The user needs the Dataplex Viewer role (roles/dataplex.viewer) — or a broader role that includes dataplex.lakes.get/list — to see the lake in the console.

Answer

They need to refresh their browser

Answer

They need the 'Owner' role

Answer

They need to be added to the project's whitelist

Question 9

Which THREE are features of Pub/Sub?

Accepted Answer

Push subscriptions. Option B (Push subscriptions) is correct because Pub/Sub supports push subscriptions, in which the service actively delivers messages to a subscriber's HTTPS endpoint as they arrive. Option C (Global scale) is correct because Pub/Sub is a fully managed, globally distributed messaging service that automatically scales to handle high throughput across regions. Option E (Pull subscriptions) is correct because Pub/Sub also supports pull subscriptions, where the subscriber client requests or streams messages from the subscription. Option A (SQL-based queries) is not a Pub/Sub feature, as Pub/Sub is a messaging service rather than a query engine, and Option D (Built-in file storage) is not provided by Pub/Sub, which stores messages transiently for delivery rather than offering file storage.

Answer

SQL-based queries

Answer

Built-in file storage

Question 10

You need to transfer files from an Amazon S3 bucket to GCS. What is the most efficient method?

Accepted Answer

Storage Transfer Service. Storage Transfer Service is a fully managed Google Cloud service designed for high-performance, scalable transfers from Amazon S3 to Cloud Storage. It handles scheduling, retries, and data integrity checks, making it the most efficient method for this use case.

Answer

Dataflow

Answer

gsutil rsync

Answer

Write a Python script

Question 11

You have a Looker dashboard that loads slowly. What is the first thing you should check in the 'Looker System Activity' or 'Query History'?

Accepted Answer

Check the Query History to see execution time and SQL complexity.. When a Looker dashboard loads slowly, the first diagnostic step is to inspect Query History (or System Activity) to see which queries are slow, how long they executed, and how complex the generated SQL is. This reveals whether the bottleneck is an expensive query, a missing index, a large data scan, or a caching issue before you change any configuration.

Answer

Restart the Looker instance.

Answer

Delete all existing users.

Answer

Change the Looker instance color theme.

Question 12

You are designing a streaming pipeline in Dataflow. You need to ensure that data is processed in the order it was generated, even if it arrives late. What do you use?

Accepted Answer

Watermarks and Allowed Lateness. Watermarks track the progress of event time in a streaming pipeline, and Allowed Lateness defines how long after the watermark passes a window will still accept late-arriving elements. Together they let Dataflow emit correct, event-time-ordered results while tolerating out-of-order arrivals. Without allowed lateness, late data is dropped once the watermark advances past the window.

Answer

Processing time triggers

Answer

Dataflow shuffle service

Answer

Fixed windows only

Question 13

When reviewing a Looker dashboard, which icon do you click to reload the data?

Accepted Answer

The refresh icon.. The refresh icon in Looker is specifically designed to reload the data displayed on a dashboard or tile. Clicking it triggers a re-execution of the underlying queries, fetching the latest data from the database. This is the standard way to update the dashboard with current information without altering its configuration.

Answer

The edit icon.

Answer

The share icon.

Answer

The filter icon.

Question 14

You are building a Looker model and want to create a measure that calculates the count of distinct customers. Which type of measure should you use?

Accepted Answer

type: count_distinct. Looker's measure type 'count_distinct' computes the number of unique values in a given dimension or column, which is exactly what is needed to count distinct customers. It is the standard LookML measure type for distinct counts.

Answer

type: count

Answer

type: average_distinct

Answer

type: sum_distinct

Question 15

Which Cloud Composer metric should you monitor to identify if your Airflow scheduler is struggling to parse your DAGs in a timely manner?

Accepted Answer

dag_processing_total_parse_time. The dag_processing_total_parse_time metric measures the cumulative time Airflow spends parsing DAG files, which directly indicates whether the scheduler is struggling to keep up with DAG parsing. A rising or consistently high value signals that DAGs are too complex, too numerous, or inefficiently written, delaying task scheduling. This is the canonical Cloud Composer metric for diagnosing scheduler parse bottlenecks.

Answer

airflow_task_execution_time

Answer

dag_run_success_count

Answer

scheduler_heartbeat