GCP-ADP Practice Test 19 — 15 Questions

Question 1

Which THREE features does Looker offer to support data governance?

Accepted Answer

Centralized LookML modeling. Looker supports data governance through centralized LookML modeling (C), because defining dimensions, measures, and business logic once in LookML creates a single source of truth that enforces consistent metric definitions across all Explores and dashboards. Content access controls (D) are correct because Looker provides role-based permissions, groups, and folder/content-level access settings that restrict who can view, edit, or manage data and dashboards. Audit logs for tracking activity (E) are correct because Looker's audit and system activity logs record user actions, queries, and content access, which is essential for compliance, monitoring, and accountability. Automated email subject line generation (A) and custom chart color themes (B) are merely presentation or scheduling conveniences and have no bearing on governing data definitions, access, or usage tracking.

Answer

Automated email subject line generation

Answer

Custom chart color themes

Question 2

You are building a Dataflow pipeline that reads from Pub/Sub and writes to BigQuery. The pipeline must handle late-arriving data up to 1 hour after the event time and still include it in the correct window. Which Dataflow feature should you configure?

Accepted Answer

Set the allowed lateness to 1 hour. Allowed lateness in Dataflow defines the additional time after the watermark that late data is accepted and processed into its appropriate window. Setting it to 1 hour ensures that events arriving up to an hour late are still included in the correct window, satisfying the requirement.

Answer

Configure the pipeline to use processing time instead of event time

Answer

Use a session window with a gap duration of 1 hour

Answer

Increase the Pub/Sub acknowledgment deadline to 1 hour

Question 3

Which THREE items can be added to a Looker Studio report to enhance interactivity?

Accepted Answer

Drop-down filter. A drop-down filter is correct because it is an interactive control that lets report viewers select one or more dimension values to filter the charts and tables on the page. A slider control is correct because it provides an interactive numeric range selector that viewers can drag to filter data bound to a metric or parameter. A date range control is correct because it lets viewers interactively change the time period applied to the report's charts, driving the displayed data. The company logo and text block are not correct because they are static presentation elements used for branding or descriptive copy and do not provide any interactive filtering or data selection capability.

Answer

Company logo

Answer

Text block

Question 4

You need to anonymize PII (Personally Identifiable Information) before loading data into BigQuery. Which tool is best?

Accepted Answer

Cloud DLP. Cloud DLP (Data Loss Prevention) is purpose-built for discovering, classifying, and de-identifying sensitive data such as PII. It provides de-identification techniques like masking, tokenization, and format-preserving encryption that can be applied to data before it lands in BigQuery. Unlike general-purpose data processing tools, Cloud DLP natively understands PII types (e.g., names, credit card numbers, SSNs) and offers configurable infoType detectors, making it the most efficient and compliant choice for anonymization.

Answer

Dataflow

Answer

BigQuery

Answer

Cloud Storage

Question 5

A compliance officer requires that a Cloud Storage bucket holding customer records retain every object version for seven years, and that no user, including project owners, be able to delete or overwrite those versions during that period. You need a configuration that enforces this even against accidental or malicious deletion attempts. What should you configure?

Accepted Answer

Apply a bucket lock on a retention policy that sets the retention period to seven years.. A locked retention policy is the only Cloud Storage mechanism that guarantees objects cannot be deleted or overwritten for a fixed period regardless of IAM permissions, because locking removes the ability to shorten or delete the policy itself. Object Versioning and role restrictions provide recoverability or reduced exposure but remain reversible, so they cannot satisfy an enforced seven-year immutability mandate.

Answer

Grant the compliance team the Storage Object Viewer role and remove Storage Object Admin from all other users.

Answer

Enable Object Versioning on the bucket and set a lifecycle rule to delete noncurrent versions after 2555 days.

Answer

Create a Cloud Storage retention policy with a seven-year period but leave it unlocked.

Question 6

A junior analyst has written a BigQuery SQL statement that joins an orders table to a customers table and then aggregates total spend per customer. Before running it against production data, the analyst wants to confirm the query is syntactically valid and to see estimated bytes processed without incurring charges for scanning the full tables. What should the analyst do?

Accepted Answer

Run the query with the BigQuery dry run option enabled to validate syntax and view the estimated bytes processed.. A dry run in BigQuery performs full SQL validation, including table and column resolution, and returns the estimated bytes processed without executing the statement or billing for data scanned. That makes it the appropriate pre-flight check for syntax and cost. Sampling data, inspecting job history, or building a materialized view all require execution or answer a different question.

Answer

Query the INFORMATION_SCHEMA.JOBS view to inspect prior executions of similar queries and their bytes billed.

Answer

Create a materialized view over the join and inspect its refresh statistics before querying it.

Answer

Run the query against a small sample table created with SELECT * LIMIT 100 to approximate the result.

Question 7

You are orchestrating a pipeline with Workflows that invokes several Cloud Functions. You need to pass the output of one function as an argument to the next. What is the correct syntax?

Accepted Answer

Assign the call result to a variable and reference it in the next step.. Workflows uses a YAML/JSON syntax where each step's return value can be assigned to a variable, and subsequent steps reference that variable using ${variableName} syntax. This is the idiomatic way to chain Cloud Functions and pass outputs forward. For example: 'call: http.post' with 'result: fn1Result', then the next step uses '${fn1Result.body}'.

Answer

Use global environment variables to store intermediate state.

Answer

Combine all function calls into a single synchronous block.

Answer

Write the output to a temporary GCS file and read it back.

Question 8

You are migrating a high-throughput on-premises database to BigQuery. You need to ensure zero downtime. What is your best strategy?

Accepted Answer

Datastream. Datastream is a serverless change data capture (CDC) service that replicates data from on-premises databases to BigQuery with minimal downtime. It continuously streams changes, ensuring zero downtime during migration.

Answer

Cloud Pub/Sub

Answer

Dataflow batch jobs

Answer

Batch Export/Import

Question 9

You are using Cloud Composer to orchestrate a data pipeline. A critical task in your DAG is failing intermittently due to a known transient error in an external API. You want to automatically retry the task up to 5 times with a 2-minute delay between retries. Which Airflow feature should you configure to achieve this?

Accepted Answer

Set `retries=5` and `retry_delay=timedelta(minutes=2)` in the task's default arguments.. Airflow's built-in retry mechanism is configured via the `retries` and `retry_delay` parameters. These can be set at the DAG level (default_args) or per task. When a task fails, Airflow will automatically retry it after the specified delay, up to the maximum number of retries. This is the standard way to handle transient failures.

Answer

Configure a `SLA` (Service Level Agreement) with a timeout to force a retry.

Answer

Implement a custom operator that loops until success, with a sleep between attempts.

Answer

Use the `on_failure_callback` to trigger a new task run after a delay.

Question 10

You need to create a visualization in Looker Studio that shows the proportion of total sales by product category. Which chart type is most appropriate?

Accepted Answer

Pie chart. A pie chart is specifically designed to show parts of a whole, with each slice representing a category's contribution to the total. When configured with product category as the dimension and sales as the metric, it visually communicates the proportion of total sales for each category, which is exactly what the scenario requires.

Answer

Scorecard

Answer

Time series chart

Answer

Geo chart

Question 11

Your organization requires that all data ingested into Cloud Storage be encrypted at rest using keys managed by you, not Google. Which feature should you implement?

Accepted Answer

Customer-Managed Encryption Keys (CMEK). Customer-Managed Encryption Keys (CMEK) in Cloud Storage let you supply your own Cloud KMS key to encrypt objects at rest, so Google does not control the key material. This satisfies the requirement that keys be managed by the organization rather than Google, and it integrates with Cloud KMS for rotation and audit.

Answer

Bucket Lock

Answer

Signed URLs

Answer

Cloud Storage default encryption

Question 12

Your team stores customer feedback in a BigQuery table with a column `feedback_text`. You need to count how many times each of the words 'fast', 'slow', and 'broken' appears across all feedback rows. You want to run this as a single SQL query in BigQuery. Which approach should you use?

Accepted Answer

Use the `REGEXP_EXTRACT_ALL` function with a pattern that matches each target word, then unnest the resulting array and group by the extracted word.. To count every occurrence of specific words across all feedback rows, you need a method that extracts each match individually and then aggregates. REGEXP_EXTRACT_ALL with a pattern matching any of the target words returns an array of all matches per row; unnesting and grouping by the extracted word gives the total count per word. This handles multiple occurrences per row and punctuation correctly.

Answer

Use the `ML.NGRAMS` function from BigQuery ML to tokenize the text and then count the n-grams.

Answer

Use the `SPLIT` function on each feedback_text row with a space delimiter, then use `UNNEST` and filter with a `WHERE` clause for the three words.

Answer

Use `COUNTIF` with three separate conditions in a single SELECT statement, such as `COUNTIF(feedback_text LIKE '%fast%')`.

Question 13

Which console menu would you navigate to in order to access Dataplex?

Accepted Answer

Analytics. Dataplex is accessed from the Analytics section of the Google Cloud Console, alongside BigQuery, Dataflow, Dataproc, and other data analytics services. This is where Google groups its data governance and analytics offerings.

Answer

Compute Engine

Answer

Operations

Answer

Storage

Question 14

A financial analyst needs a Looker Studio report that displays the same revenue chart but filtered by a different region on each of four pages, without duplicating the underlying BigQuery data source. The report currently uses one chart that reads from a BigQuery view. What should the analyst do to reuse the chart across pages with different region filters?

Accepted Answer

Create a reusable chart as a component in the report's component library, then place it on each page and bind a page-specific region filter.. Reusable components let a single chart definition be placed on several pages while each placement keeps its own page-level filter controls. Because the component stays bound to the same BigQuery data source, revenue logic is maintained once and each page can target a different region. Duplicating views or relying on report-level filters either multiplies maintenance or forces one filter across all pages.

Answer

Create four separate BigQuery views, one per region, and add a chart for each view on its own page.

Answer

Add a report-level filter for region and let each page inherit it, then rely on page-level filter overrides.

Answer

Use a blend between the BigQuery view and a manually entered region parameter table, then add a parameter control per page.

Question 15

You need to load a 5 TB uncompressed CSV file from Cloud Storage into BigQuery. The file is too large to load in a single job. What is the recommended approach?

Accepted Answer

Split the file into multiple smaller files and load them in parallel using multiple load jobs.. For loading very large files into BigQuery, splitting the data into multiple smaller files and using parallel load jobs is the most efficient and reliable method. It avoids per-job size limits and leverages BigQuery's parallel load capabilities. Other methods like streaming or external tables are not optimal for bulk loading.

Answer

Create an external table in BigQuery that points to the CSV file in Cloud Storage.

Answer

Compress the file with gzip and load it as a single file.

Answer

Use the BigQuery Storage Write API to stream the file directly into BigQuery.