Ingest and Transform Data DP-203 Practice Questions (Page 8)

Questions 36–40

1. 36foundation · easy

   What is the purpose of configuring error handling in a data transformation pipeline?

   • ATo encrypt error messages for security
   • BTo prevent any errors from ever occurring
   • CTo automatically fix all data quality issues
   • DTo define how the pipeline responds when a transformation fails
   Reveal Solution Discussion Select an answer first
2. 37application · medium

   A security analytics company ingests log files that contain base64-encoded payloads. The payloads need to be decoded before further analysis. You are using Apache Spark in Azure Synapse Analytics. What should you do?

   • AUse the decode function with the base64 charset on the payload column.
   • BUse the unbase64 function on the payload column.
   • CUse the cast function to convert the payload column to binary.
   • DUse the from_json function to parse the payload as JSON.
   Reveal Solution Discussion Select an answer first
4. 38foundation · easy

   What does 'Exactly Once Delivery' mean in the context of Azure Stream Analytics?

   • AEvents are delivered to multiple outputs for redundancy
   • BEvents are delivered only when the system is idle
   • CEvents are delivered at most once, allowing some loss
   • DEach event is processed exactly once, preventing duplicates and avoiding data loss
   Reveal Solution Discussion Select an answer first
5. 39expert · hard

   A telemetry company uses Azure Stream Analytics to process IoT sensor data. The data is generated by devices that occasionally go offline and then send a batch of historical data when they reconnect. The company needs to ensure that this late-arriving data is processed correctly, but also wants to avoid processing data that is too old. You need to configure the Stream Analytics job to handle this. What should you do?

   • AUse a tumbling window with a late-arrival policy that allows a grace period of 5 minutes.
   • BUse a hopping window with a large hop size to capture all late data.
   • CUse a session window with a timeout to group late-arriving data.
   • DUse a sliding window with a long duration to include all late data.
   Reveal Solution Discussion Select an answer first
6. 40expert · hard

   A telemetry company ingests IoT sensor data into Azure Event Hubs. The data is processed by an Azure Stream Analytics job that writes results to Azure Synapse Analytics. The job must not produce duplicate output records even if the input is delivered more than once. Additionally, some events arrive out of order due to network delays. You need to configure the Stream Analytics job to handle both requirements. What should you do?

   • AEnable Exactly Once Delivery on the output and use a Pass-Through query with no time window.
   • BConfigure the job with a Hopping Window and set the output to 'Exactly Once' delivery.
   • CUse a Tumbling Window with a large window size and rely on the output's built-in deduplication.
   • DSet the late-arrival policy to 'Adjust' and enable Exactly Once Delivery on the output.
   Reveal Solution Discussion Select an answer first