Task 2.4: Design Data Models and Schema Evolution DEA-C01 Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A data science team is developing a new machine learning model in Amazon SageMaker. They need to document the lineage of the model, including the specific dataset version used for training, the preprocessing code, and the model artifacts. This is required for internal audit and reproducibility. Which AWS service or feature should they use to automatically capture and query this lineage information?

   • AUse Amazon S3 server access logs to track which files were read during training.
   • BUse AWS CloudTrail to log all SageMaker API calls.
   • CUse AWS Glue DataBrew to profile the data and track its lineage.
   • DUse Amazon SageMaker ML Lineage Tracking to record the relationships between datasets, code, and model artifacts.
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   Which compression technique is commonly used in columnar data stores like Amazon Redshift to reduce storage and improve query performance?

   • ARow-level compression
   • BDictionary compression only
   • CIndex compression
   • DColumn encoding
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A company is building a semantic search application using Amazon Bedrock Knowledge Bases. They have a large corpus of legal documents. The application needs to return the most relevant documents based on a user's natural language query. The team has decided to use a vector store to enable this functionality. What is the primary role of the vector store in this architecture?

   • ATo store the embeddings of the documents and enable similarity search based on vector distance.
   • BTo store the original text of the documents and use SQL to search for keywords.
   • CTo store the documents in a columnar format and use compression to reduce storage.
   • DTo store the documents in a key-value store and use the key to retrieve the document.
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   What is a common trade-off when adding too many indexes to a table?

   • AIncreased write latency and storage overhead
   • BAutomatic data compression
   • CReduced data durability
   • DDecreased read performance for all queries
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A company stores IoT sensor data in an Amazon S3 data lake. Initially, data was written in small batches every hour. Now, the sensors are sending data continuously, and the data volume has increased 100x. The current schema uses a single table with a partition per day. Queries are becoming slow and expensive. Which schema evolution strategy is most appropriate to address the change in data velocity and volume?

   • AKeep the daily partition structure but switch the storage format to CSV.
   • BCreate a new table for each sensor to isolate the data.
   • CRemove all partitioning and store data in a single large file.
   • DAdd a partition for each hour to the existing table to reduce the data scanned per query.
   Reveal Solution Discussion Select an answer first