Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Associate Data Practitioner practice bank has 1,132 more like it. Try 10 free: instant score, no account.
A retail company is designing a new data pipeline to process high volumes of daily sales logs from multiple regional branches. The data engineering team has established the following architectural requirements:
Which data integration pattern and approach best fulfills these requirements?
ELT (Extract, Load, Transform) is a modern data integration architecture where raw data is extracted from source systems, immediately loaded into a centralized data warehouse or data lake, and subsequently transformed within the target storage engine using its native compute capabilities.
Unlike traditional ETL, which requires transforming data in-flight before it reaches the warehouse, ELT takes full advantage of cloud-native data warehouses like BigQuery. It decouples the loading process from data modeling, allowing analysts and data engineers to iterate on transformations rapidly using standard SQL.
Reverse ETL is a data integration pattern where processed, modeled data from a centralized data warehouse (like BigQuery) is copied back into operational systems, CRMs, or transactional databases (such as Salesforce, HubSpot, or operational datastores) to drive business workflows.
The team already had transformed, aggregated customer segment metrics in BigQuery and needed to automatically sync those calculated scores back into an operational application or CRM database.
This approach uses an ETL (Extract, Transform, Load) design where a managed pipeline service like Cloud Data Fusion extracts source data, cleanses and formats it on an intermediate compute layer, and then loads only the final transformed data into a transactional database such as Cloud SQL.
The organization needed to sanitize sensitive personally identifiable information (PII) before data entered any cloud environment, or if the destination was an operational relational database requiring strict pre-load schema validation.
This approach describes a pre-load ETL pattern where raw files are parsed and transformed in-flight using Dataflow (Apache Beam) to perform transformations and aggregations before persisting the final output files into Cloud Storage.
The scenario required real-time streaming transformations with sub-second latency from Pub/Sub, or complex sessionization and enrichment that cannot be expressed easily in standard SQL.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated