Develop a Batch Processing Solution DP-203 Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   In Azure Data Factory, which policy on an activity allows you to specify the number of times to retry an activity if it fails, as part of exception handling?

   • ARetry policy
   • BConcurrency policy
   • CDependency policy
   • DTimeout policy
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A healthcare organization stores patient data in Azure Data Lake Storage Gen2. The data engineering team is designing a batch processing solution that reads from the raw zone, transforms the data, and writes to the curated zone. The team needs to ensure that if a transformation job fails, the raw data is not corrupted and the job can be retried safely. They also need to retain the raw data for 7 years for compliance. What should the team implement?

   • AConfigure a lifecycle management policy to delete the raw data after 30 days and keep only the curated data.
   • BWrite transformation results to a temporary directory in the curated zone and atomically rename the directory to the final location after success.
   • CUse a 'COPY INTO' command to load the raw data into the curated zone, which automatically handles retries and failures.
   • DOverwrite the raw data with the transformed data after a successful job to save storage costs.
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Which Azure service provides a collaborative Apache Spark–based analytics platform that can be used to develop and run batch processing jobs in Python, Scala, or SQL?

   • AAzure Synapse Analytics serverless SQL pool
   • BAzure Databricks
   • CAzure Batch
   • DAzure Data Factory
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A media company uses Azure Databricks to process clickstream data stored in Delta Lake. The batch job reads new events, enriches them with user profile data, and writes the results back to Delta Lake. The job currently processes data in large partitions, causing frequent out-of-memory errors on the executors. The team needs to improve stability without changing the cluster configuration. What should the team modify in the notebook?

   • ASet the 'spark.sql.autoBroadcastJoinThreshold' to a higher value to optimize the join operation.
   • BUse 'repartition' to increase the number of partitions in the DataFrame before performing the enrichment and write operations.
   • CSwitch to 'coalesce' to reduce the number of partitions and decrease the overhead of task scheduling.
   • DIncrease the number of partitions when reading the source data by using a larger value for the 'spark.sql.shuffle.partitions' setting.
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A healthcare company wants to run near-real-time analytics on data stored in an Azure SQL Database. They currently use a nightly batch process to copy data into Azure Synapse Analytics. The business now needs fresher data for reporting, but they want to avoid building a new ETL pipeline. What should you implement?

   • AUse PolyBase to create an external table that queries the SQL Database directly.
   • BUse Azure Synapse Link to replicate the SQL Database data into the dedicated SQL pool automatically.
   • CUse Azure Databricks to read from the SQL Database and write to the dedicated SQL pool every few minutes.
   • DUse Azure Data Factory to run the existing pipeline every 15 minutes instead of nightly.
   Reveal Solution Discussion Select an answer first