Design and Implement Loading Patterns DP-700 Practice Questions (Page 7)

Questions 31–35

1. 31application · medium

   A logistics company needs to ingest real-time GPS location data from thousands of vehicles into Fabric for live tracking dashboards. The data arrives as JSON messages through an event hub. The solution must support near-real-time analytics with minimal latency. What should you implement?

   • AA scheduled Dataflow Gen2 that runs every 15 minutes to read from the event hub and write to a Lakehouse table.
   • BA Pipeline that copies the event hub data to a Lakehouse table once per hour.
   • CA Spark structured streaming job that reads from the event hub and writes aggregated data to a Lakehouse table in Delta format.
   • DA SQL query that directly queries the event hub using T-SQL and returns results to the dashboard.
   Reveal Solution Discussion Select an answer first
2. 32expert · hard

   A manufacturing company loads machine telemetry data into a Fabric warehouse every hour. The source system provides a full export of all telemetry data each hour. The team wants to reduce the data volume loaded and only load new or changed data. The source does not have a last-modified timestamp. What should they do?

   • AUse a Spark notebook to read the full export, compare it with the existing warehouse table, and write only the new or changed rows.
   • BUse a Dataflow Gen2 to compare the full export with the warehouse table and load only the differences.
   • CLoad the full export into a staging table and use a MERGE to update the warehouse table.
   • DLoad the full export into the warehouse table and use a view to filter out unchanged rows.
   Reveal Solution Discussion Select an answer first
4. 33expert · hard

   A global bank is migrating its historical transaction data to Fabric. The data is stored in multiple on-premises databases and totals 10 TB. The migration must complete within a 3-day weekend. The bank has a 1 Gbps network connection to Azure. The data must be cleansed and transformed to fit a star schema. What should you do to meet the deadline?

   • AUse a Dataflow Gen2 to copy the data directly from the on-premises databases to Fabric tables without transformation.
   • BUse Azure Data Factory pipelines to copy the raw data to Azure Blob Storage, then use Spark notebooks in Fabric to transform and load the data into the star schema.
   • CUse a Spark notebook to read directly from the on-premises databases over the network and write to Fabric tables.
   • DUse a VPN connection to mount the on-premises databases as external tables in Fabric and query them directly.
   Reveal Solution Discussion Select an answer first
5. 34application · medium

   A data engineer is loading sales data into a star schema in a Fabric warehouse. The source has a single table with customer names, product names, and sales amounts. The engineer needs to populate the dimension tables (Customer and Product) and the fact table (Sales). What should the engineer do?

   • ACreate a single table with all columns and use it for both dimensions and facts.
   • BLoad the source table directly into the fact table and create views for dimensions.
   • CUse a notebook to transform the source into separate dimension and fact tables with surrogate keys.
   • DLoad the source into the warehouse and use Power BI to create the dimensions and facts.
   Reveal Solution Discussion Select an answer first
6. 35expert · hard

   A social media company needs to ingest user interaction events into Fabric for real-time analytics. The events are sent to an event hub. The team wants to use a serverless approach to minimize cost when there is no traffic. What should they implement?

   • AA Spark structured streaming job that runs continuously with a 1-second trigger.
   • BA Pipeline that runs on a schedule to copy events from the event hub to a Lakehouse table.
   • CA scheduled Dataflow Gen2 that runs every 5 minutes to process events from the event hub.
   • DA Fabric eventstream that ingests events from the event hub and writes them to a Lakehouse table, with autoscaling based on traffic.
   Reveal Solution Discussion Select an answer first