Build Data Pipelines for Data Science Workloads SNOWPRO-ADVANCED-DATA-SCIENTIST Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   What is the primary benefit of using clustering keys in a Snowflake table?

   • AReduced storage costs
   • BImproved query performance on clustered columns
   • CEnhanced security
   • DAutomatic data cleaning
   Reveal Solution Discussion Select an answer first
2. 22foundation · easy

   Which Snowflake object is used to schedule SQL statements to run automatically at specified intervals?

   • AStage
   • BView
   • CTask
   • DStream
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   What is a common technique to handle outliers in a dataset during pipeline processing?

   • AImputing missing values with the mean
   • BClipping or winsorizing extreme values
   • COne-hot encoding categorical variables
   • DNormalizing the data to a range of 0-1
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   How can Snowflake integrate with external ML frameworks for model training?

   • ABy using tasks to run external ML jobs
   • BBy using external functions to call ML APIs
   • CBy using streams to send data to external frameworks
   • DBy using views to expose data to external frameworks
   Reveal Solution Discussion Select an answer first
6. 25expert · hard

   A data science team is building a feature engineering pipeline that creates features from raw event data. The features are used for multiple models, and the team wants to ensure that feature definitions are consistent across models and that feature values are versioned. They also want to automate the pipeline to run daily. Which approach should they use?

   • ACreate a view for each feature and use it directly in models.
   • BUse Snowpark to compute features on the fly for each model training run.
   • CUse a stored procedure to compute features and store them in a table without versioning.
   • DCreate a feature store table that stores feature values with a version column, and use a task to populate it daily.
   Reveal Solution Discussion Select an answer first