Implement a Partition Strategy DP-203 Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A media streaming company stores user activity logs in Azure Data Lake Storage Gen2. Data scientists run exploratory queries that filter by user_id and by date. The current layout is /logs/2024/05/15/activity.parquet. Queries are slow. You need to determine whether partitioning is necessary and, if so, implement it. What should you do?

   • APartition only by user_id, ignoring date.
   • BKeep the current layout because partitioning is not needed for exploratory queries.
   • CConvert files to CSV and use Azure Synapse Serverless SQL.
   • DPartition by user_id and date, e.g., /logs/user_id=123/date=2024-05-15/activity.parquet.
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   When is it most appropriate to implement partitioning in Azure Data Lake Storage Gen2?

   • AWhen queries frequently filter by a specific column and the dataset is large
   • BWhen the data is already compressed and stored in Parquet format
   • CWhen the data lake contains a small number of large files that are rarely queried
   • DWhen the data is stored in a single file and is accessed by a single application
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A financial services company stores transaction data in Azure Data Lake Storage Gen2. The data is partitioned by date and then by region. Each day, about 10,000 small files are written. Data scientists run queries that filter on a specific account ID and a date range of one month. Queries are slow because they scan many small files. You need to improve query performance while minimizing changes to the ingestion pipeline. What should you do?

   • AChange the partition structure to region first, then date.
   • BUse Azure Data Factory to compact small files into larger Parquet files after ingestion.
   • CEnable Azure Data Lake Storage Gen2 hierarchical namespace.
   • DAdd an account ID level to the folder structure: /transactions/{accountId}/{date}/{region}/
   Reveal Solution Discussion Select an answer first
5. 9expert · hard

   A media streaming company ingests viewer activity events into Azure Event Hubs. Each event includes a user ID, a content ID, and a timestamp. Downstream Azure Stream Analytics aggregates events by user and by 1-minute tumbling windows. The current partition key is the content ID. You observe that popular content causes partition skew, leading to processing delays. You need to reduce skew while preserving per-user ordering for windowed aggregations. What should you do?

   • AChange the partition key to user ID.
   • BIncrease the partition count by recreating the Event Hub.
   • CUse a composite key of user ID and content ID.
   • DUse a random partition key to distribute evenly.
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A telecom company uses Azure Synapse dedicated SQL pool for a call detail record (CDR) fact table. The table is partitioned by month and distributed by hash on PhoneNumber. Analysts run queries that filter on a specific cell tower ID and a date range of one day. You notice that queries scan multiple partitions and performance is poor. What is the best way to improve query performance?

   • ARe-partition the table by day and keep cell tower ID as a secondary partition.
   • BChange the partition column to cell tower ID and keep monthly partitions.
   • CCreate statistics on cell tower ID and date columns.
   • DUse a round-robin distribution and partition by cell tower ID.
   Reveal Solution Discussion Select an answer first