MLA-C01 · Question #137
A company is using Amazon EMR. The company has a large dataset in Amazon S3 that needs to be ingested into Amazon SageMaker Feature Store. The dataset contains historical data and real-time…
The correct answer is C. Use the Feature Store Spark connector to ingest the data as Spark DataFrames with the online. Option C is correct because the company already uses Amazon EMR (a Spark environment), and the Feature Store Spark connector is specifically designed for bulk ingestion of large datasets as Spark DataFrames. When configured to target both the online and offline stores…
Question
A company is using Amazon EMR. The company has a large dataset in Amazon S3 that needs to be ingested into Amazon SageMaker Feature Store. The dataset contains historical data and real-time streaming data. The company must ensure that the Feature Store online store is updated with the most recent data as soon as the data becomes available. The company also must maintain a complete Feature Store offline store for batch processing. Which solution will meet these requirements?
Options
- AUse the PutRecord API in Feature Store Runtime to ingest all the data into the online store.
- BUse the PutRecord API in Feature Store Runtime to ingest all the data into the offline store.
- CUse the Feature Store Spark connector to ingest the data as Spark DataFrames with the online
- DUse the Feature Store Spark connector to ingest the data as Spark DataFrames with only the
How the community answered
(56 responses)- A14% (8)
- B4% (2)
- C75% (42)
- D7% (4)
Explanation
Option C is correct because the company already uses Amazon EMR (a Spark environment), and the Feature Store Spark connector is specifically designed for bulk ingestion of large datasets as Spark DataFrames. When configured to target both the online and offline stores simultaneously, it satisfies both requirements in a single operation: the online store receives the latest record values immediately for low-latency lookups, while the offline store retains a complete history suitable for batch processing and model training.
Why the distractors are wrong:
- A (PutRecord → online store only): PutRecord is a record-by-record API-inefficient for large datasets-and writing only to the online store discards the historical completeness needed for the offline batch requirement.
- B (PutRecord → offline store only): Omits the online store entirely, so real-time feature serving would have no up-to-date data; also still suffers from the single-record inefficiency at scale.
- D (Spark connector → online store only): Leverages the right tool for scale but fails to maintain the complete offline store history required for batch processing.
Memory tip: Match the ingestion tool to the job - Spark connector = bulk/EMR, PutRecord = single records/streaming. When a question mentions EMR + large dataset + both stores needed, the Spark connector with both online and offline targets is always the right pairing.
Topics
Community Discussion
No community discussion yet for this question.