Design and Implement Loading Patterns DP-700 Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   When implementing a full load of a large dataset, which practice helps improve performance?

   • AUsing a single large file to reduce overhead
   • BDisabling error handling to speed up the load
   • CLoading data in parallel from multiple files
   • DLoading data row by row using a loop
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   Which component is commonly used in an incremental load design to track the last successfully processed record?

   • AWatermark column
   • BForeign key constraint
   • CPrimary key index
   • DDefault constraint
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A healthcare company receives daily patient admission files from hospitals. Each file contains new admissions and updates to existing admissions. The team needs to maintain a history of all admissions, including changes over time. The engineer must design a loading pattern that tracks changes without losing historical data. What should the engineer do?

   • AAppend all rows from the daily files to the table, including updates as new rows.
   • BTruncate the table and reload all admissions each day.
   • CUse a type 2 dimension pattern with effective dates to track changes.
   • DUse a merge operation to update existing rows and insert new ones.
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A company streams financial transactions into an event hub. The data is used for real-time fraud detection and also for end-of-day reconciliation. The team needs to ensure that the real-time processing has low latency, while the reconciliation can be done in batches. The engineer must design a streaming load pattern that meets both requirements. What should the engineer do?

   • AUse a Dataflow Gen2 to ingest data every hour and load into a lakehouse.
   • BUse a scheduled notebook to ingest data every 5 minutes and process it in batches.
   • CUse a KQL database to ingest the stream and perform both real-time and batch analysis.
   • DUse Eventstream to ingest data, write raw to a lakehouse, and use a Spark streaming job for fraud detection.
   Reveal Solution Discussion Select an answer first
6. 20expert · hard

   A retail company is building a star schema in Fabric to analyze sales. The source data contains multiple fact tables and dimension tables. The team needs to load the data in a way that supports incremental updates and maintains referential integrity. What should they implement?

   • ALoad fact tables first, then generate dimension keys from the fact data.
   • BLoad all dimensions first, then load all fact tables, using surrogate keys generated during the dimension load.
   • CLoad dimensions and fact tables in parallel to save time.
   • DLoad all data into a single table to simplify the process.
   Reveal Solution Discussion Select an answer first