Data Practitioner Practice Exam 2025: Latest Questions

A retail team receives product catalog data as nested JSON with arrays (variants, attributes) and wants to run ad-hoc SQL analysis with minimal schema management. Which storage option is most appropriate?

You need to load daily CSV files from a Cloud Storage bucket into BigQuery. The files have the same schema each day, and the team wants a simple, serverless approach. What should you use?

A data analyst reports that a BigQuery query returns a different number of rows each time it runs, even though the table data is not changing. The query uses RAND() to sample records. What is the most likely explanation?

A dataset contains customer records, including a column with national ID numbers. Analysts should be able to query other fields, but only a small security group can see the national ID column. What is the recommended BigQuery access control approach?

A company wants to process streaming click events with low-latency transformations (enrichment, filtering) and write results to BigQuery for analytics. The solution should be managed and support both streaming and batch patterns. What should you use?

A team stores web logs in BigQuery partitioned by event_date. A query filtering by a timestamp column (event_ts) is slow and scans many partitions. What is the best practice to reduce data scanned?

A data engineering team needs a repeatable way to deploy BigQuery datasets, tables, and IAM bindings across dev/test/prod with peer review and change history. What approach best meets this requirement?

A data analyst wants to explore a dataset and create interactive dashboards for business users, using BigQuery as the primary data source with minimal engineering effort. Which tool is the best fit?

A regulated healthcare organization must ensure that sensitive data in BigQuery cannot be exfiltrated to the public internet, even by a user with BigQuery permissions. They also need to use private access to Google APIs. What is the best control to implement?

Your company has an existing Apache Hadoop/Spark workload that relies on HDFS semantics and needs minimal code changes while migrating to Google Cloud. The workload is batch-oriented, and the team is comfortable managing clusters but wants a managed offering. What should you choose?

You are ingesting clickstream events into BigQuery. Each event has a fixed schema (user_id, event_name, event_time) and an additional set of arbitrary key-value attributes that vary by event type. You need to keep the core schema stable while still allowing flexible attributes to be queried. What is the BEST BigQuery modeling approach?

A data analyst asks for a simple way to explore a BigQuery dataset and build a basic dashboard with shareable charts, without managing infrastructure. Which Google Cloud tool is the most appropriate?

Your team wants to share a curated BigQuery dataset across multiple projects while ensuring consumers see only the approved tables and do not get broad access to the source project. What is the recommended approach?

A Pub/Sub subscription is pushing messages to a Dataflow streaming pipeline, but you notice message duplication in BigQuery. You must reduce the impact of duplicates without sacrificing throughput. Which approach is most appropriate?

You have a BigQuery table partitioned by event_date. Analysts often filter by event_date and user_id, but queries are still scanning more data than expected. What should you do to improve performance and reduce scanned bytes?

A team needs to run the same BigQuery transformation every hour and maintain a history of results. They want a managed scheduler with minimal operational overhead. What should they use?

You are designing storage for sensor time-series data that arrives continuously. The application needs very low-latency reads of the latest measurements per device, and you also need to retain large volumes for analysis later in BigQuery. Which design is most appropriate?

A dataset contains a "customer_email" column that should not be visible to most analysts. You need to let analysts query the rest of the table while masking or restricting access to that column. What is the BEST solution in BigQuery?

You are troubleshooting a BigQuery query that unexpectedly returns inflated totals after joining two tables: orders and order_items. Each order can have multiple items. The analyst expects total revenue per day from orders, but the join multiplies order-level fields. What is the most appropriate fix?

A data engineer receives CSV files where numeric columns sometimes contain non-numeric placeholders like "N/A". They must load the data into BigQuery while preserving the original values for auditing and enabling correct numeric analysis. What should they do?

A retail team loads daily product catalogs into BigQuery. Some columns are inconsistently typed across sources (for example, "price" arrives as STRING in one file and NUMERIC in another). Analysts need stable types for dashboards. What is the best practice to handle this in BigQuery?

A data practitioner needs to explore a 20 TB BigQuery table to find suspicious purchase patterns. They want a quick summary of value distributions and correlations without writing many custom queries. Which BigQuery capability is most appropriate?

A healthcare organization stores patient events in BigQuery. Compliance requires limiting access so analysts can only see de-identified columns, while a small clinical group can see full records. Both groups must query the same underlying tables. What is the recommended approach?

A team uses Pub/Sub to ingest clickstream events and streams them into BigQuery. They notice occasional duplicate rows in BigQuery when Dataflow workers retry after transient failures. They need to minimize duplicates without sacrificing throughput. What should they do?

A company builds a shared analytics dataset in BigQuery that includes data from multiple business units. They must prevent accidental exposure of PII and ensure new columns are classified and protected before analysts can query them. What is the best design?