Implement a Partition Strategy DP-203 Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A telemetry platform ingests IoT device messages into Azure Event Hubs. Each message includes a device ID, a timestamp, and a reading. Downstream Azure Stream Analytics jobs aggregate data by device and by 5-minute tumbling windows. The current Event Hubs partition count is 8. You notice that some partitions are heavily skewed, causing processing delays. What should you do to improve throughput and reduce skew?

   • AIncrease the partition count to 32 and use device ID as the partition key.
   • BRecreate the Event Hub with more partitions and use device ID as the partition key.
   • CKeep the partition count but change the partition key to a random GUID.
   • DUse a composite partition key of device ID and timestamp.
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A healthcare analytics team uses Azure Synapse dedicated SQL pool for a fact table containing patient encounters. The table is partitioned by year and distributed by hash on PatientID. Queries often filter on a specific clinic ID and a date range of one month. The table has 10 years of data. You need to improve query performance. What should you do?

   • AChange the partition column to clinic ID and keep yearly partitions.
   • BRe-partition the table by month and keep clinic ID as a secondary partition.
   • CUse a round-robin distribution and partition by clinic ID.
   • DCreate statistics on clinic ID and date columns.
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Which file partitioning approach is most effective for managing incremental data loads in a data lake?

   • APartition by file size
   • BPartition by ingestion timestamp (e.g., year/month/day/hour)
   • CPartition by random UUID
   • DPartition by source system only
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A telemetry company ingests IoT sensor readings from millions of devices into Azure Event Hubs. The data is then written to Azure Data Lake Storage Gen2 in near real-time. Downstream analytics often filter by device_id and by hour. The current writer creates a new file every minute with all events. Queries are slow and file counts are high. You need to design a file partition strategy that balances query performance and manageable file counts. What should you do?

   • AWrite files partitioned only by device_id, ignoring time.
   • BWrite a new file every minute without any folder partitioning.
   • CWrite files partitioned by device_id and hour, e.g., /telemetry/device_id=123/hour=2024-05-15-14/data.parquet.
   • DWrite one large file per day containing all device events.
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   A healthcare company uses Azure Synapse Analytics dedicated SQL pool for its data warehouse. The fact table contains patient visits and is partitioned by visit_date. Queries that filter by visit_date are fast, but queries that filter by patient_id are slow because they scan all partitions. You need to improve performance for patient_id queries without degrading visit_date query performance. You have a limited budget for storage and maintenance. What should you do?

   • ARecreate the table with hash distribution on patient_id and keep partitioning on visit_date.
   • BCreate a clustered columnstore index on patient_id.
   • CAdd a secondary partition on patient_id to the existing table.
   • DCreate a materialized view on patient_id that aggregates the fact table.
   Reveal Solution Discussion Select an answer first