Design and Implement Data Pipelines DP-750 Practice Questions (Page 7)

Questions 31–35

1. 31expert · hard

   A team is building a notebook-based pipeline in a Lakeflow job. The pipeline has a task that reads data from a source, a task that transforms the data, and a task that writes the data to a destination. The team wants to implement error handling so that if the transform task fails, the write task does not run, but the read task should not be retried. They also want to send an alert when the transform task fails. What should they do?

   • ASet the write task to depend on the read task, and set the transform task to depend on the write task.
   • BSet the transform task to depend on the read task, and set the write task to depend on the transform task. Configure an alert on the transform task failure.
   • CCombine the read and transform tasks into a single notebook, so that if transform fails, the read task is retried.
   • DUse a Lakeflow Spark Declarative Pipeline instead of notebooks, because it provides built-in alerting on task failure.
   Reveal Solution Discussion Select an answer first
2. 32foundation · easy

   In a notebook-based data pipeline, what is the purpose of setting a precedence constraint between two notebook tasks?

   • ATo ensure the second notebook runs only after the first notebook completes successfully
   • BTo allow both notebooks to run simultaneously
   • CTo share variables between the notebooks
   • DTo combine the code of both notebooks into a single notebook
   Reveal Solution Discussion Select an answer first
4. 33foundation · easy

   Which scenario is best suited for a Lakeflow Spark Declarative Pipeline rather than a notebook-based pipeline?

   • AA pipeline that applies a series of standard transformations to incrementally load data into a Delta table
   • BA pipeline that must be developed by data scientists who are not familiar with SQL or Spark
   • CA pipeline that requires interactive exploration of data before writing any transformation logic
   • DA pipeline that needs to call external APIs and handle custom error conditions
   Reveal Solution Discussion Select an answer first
5. 34foundation · easy

   Which error handling strategy is most appropriate for a notebook that reads a file that may occasionally be missing?

   • ALet the notebook fail and rely on the job retry policy to rerun the entire pipeline
   • BUse a global variable to store the file path and change it manually each time
   • CWrap the file read in a try-except block and log a warning, then continue with an empty DataFrame
   • DIgnore the error and continue with the next cell
   Reveal Solution Discussion Select an answer first
6. 35application · medium

   A team is designing a Lakeflow job with multiple tasks. The job must first run a data ingestion task, then run a data validation task. If validation fails, the job should send an alert and stop. If validation succeeds, the job should run a data transformation task. How should the team design the task logic?

   • ACreate three tasks with no dependencies, and have each task check a shared flag to decide whether to run.
   • BCreate three tasks: ingestion, validation, and transformation. Set the validation task to depend on ingestion, and set the transformation task to depend on validation. Configure the validation task to fail the job if validation fails, and set up an alert on job failure.
   • CCreate two tasks: ingestion and a combined validation/transformation task. The combined task runs validation and then transformation, and sends an alert if validation fails.
   • DCreate an ingestion task and a transformation task, and embed the validation logic inside the transformation task, so the transformation task fails if validation fails.
   Reveal Solution Discussion Select an answer first