Associate Data Practitioner

Last checked October 22, 2026 against the official Google Cloud exam page.

1. Question 1Prepare and ingest data

   A web team wants clickstream events to appear in BigQuery dashboards within seconds of each page view. Which ingestion design meets this requirement?

   1. AFederated query that reads the application's Cloud SQL database directly
   2. BNightly batch load of CSV exports from Cloud Storage into BigQuery
   3. CScheduled query that copies the full events table every hour
   4. DPub/Sub topic with a streaming Dataflow job writing to BigQuery
   Show answer and explanationHide answer

   Correct answer: D

   Sub-minute freshness needs a streaming path, and Pub/Sub with a streaming Dataflow pipeline is the standard managed design for it. Nightly loads and hourly copies cannot meet a seconds-level requirement. The Cloud SQL federated query is the tempting wrong answer, but it reads the operational database, adds load to it, and is not a streaming ingestion pattern.

2. Question 2Analyze and present data

   A sales dashboard reads a raw events table that is partitioned by day and has years of history, so it is slow and expensive. Which TWO actions help? (Choose two.)

   1. AConnect the dashboard directly to the raw events table with no filters applied
   2. BSchedule a query that materialises a daily aggregate into a reporting table
   3. CExport the dashboard as a PDF attached to a weekly email to the sales team
   4. DCopy the table to a local spreadsheet every morning before the team starts work
   5. ECreate a view that limits results to the last 90 days and point the dashboard at it
   Show answer and explanationHide answer

   Correct answer: B, E

   A scheduled aggregate table reduces the bytes each refresh scans, and a view restricted to a recent window stops the dashboard from reading old partitions. Connecting to the raw table with no filters maximises scan cost, and the PDF and spreadsheet options do not change the query at all.

3. Question 3Manage data

   An analytics table holds customer email addresses next to country and order count. Analysts need country-level counts but must not see emails. Which control is most appropriate?

   1. AGrant analysts Viewer on the dataset and ask them not to select the email column
   2. BApply a policy tag to the email column and restrict it to a privacy group
   3. CEncrypt the email column with a customer-managed key and share the key with analysts
   4. DExport nightly copies with the email column removed and give analysts those copies
   Show answer and explanationHide answer

   Correct answer: B

   A policy tag enforces column-level access inside BigQuery, so analysts who lack the Fine-Grained Reader role cannot read the email values even when they query the table. Asking analysts to behave is not a control, and a stripped export creates a second copy that must be maintained. Encryption with a shared key does not restrict who can read the column.

4. Question 4Orchestrate data pipelines

   A daily pipeline must start only after a vendor file lands in a Cloud Storage bucket, retry failed steps, and run its steps in a defined order. Which option best fits?

   1. AA Cloud Composer DAG that uses a GCS object sensor and task retries
   2. BA cron job on a Compute Engine VM that runs a shell loop over the bucket
   3. CA Cloud Scheduler job that triggers a Cloud Function every five minutes
   4. DA notebook that a practitioner runs manually after checking the bucket
   Show answer and explanationHide answer

   Correct answer: A

   Cloud Composer, which is managed Apache Airflow, gives you dependencies between tasks, a sensor that waits for the file, and per-task retries with logging. The VM cron loop has none of the dependency or retry handling, and a five-minute Cloud Scheduler trigger keeps firing with no notion of which steps already succeeded. The manual notebook depends on a person being available.

5. Question 5Prepare and ingest data

   A partner's CSV file stores some dates as 03/04/2025 and others as 2025-04-03. The target BigQuery column is a DATE type. What should the practitioner do first?

   1. AStandardise dates to ISO 8601 in a transform step before the load
   2. BLoad the column as STRING and let each analyst cast it in their own query
   3. CSet max_bad_records to a high number so the load job always completes
   4. DEnable allow_jagged_rows on the load job so malformed dates are skipped
   Show answer and explanationHide answer

   Correct answer: A

   Standardising to ISO 8601 before the load makes the column consistent and loadable into a DATE type, so every downstream query behaves the same way. Loading as STRING pushes the cleanup onto every analyst and invites inconsistent results. Raising max_bad_records only hides rejected rows, and allow_jagged_rows concerns rows with missing or extra columns, not date formats.