Which solution will meet these requirements?
Use the SageMaker Feature Store BatchGetRecord API with the record identifier. Filter to find the latest record.
Create an Amazon Athena query to retrieve the data from the feature table.
Create an Amazon Athena query to retrieve the data from the feature table. Use the write_time value to find the latest record.
Use the SageMaker Feature Store GetRecord API with the record identifier.
Explanations:
The BatchGetRecord API is designed for batch processing and retrieves multiple records at once. It does not support fetching only the latest version of a single record.
While Amazon Athena can query the feature store, it is not ideal for real-time inference because it is more suited for batch analysis or querying historical data, not for retrieving the latest customer metadata in real-time.
Athena is a good tool for querying data in SageMaker Feature Store, but using it in this case would involve unnecessary complexity and latency for real-time inference. Athena isn’t designed for real-time retrieval of the latest record.
The SageMaker Feature Store GetRecord API retrieves a single record, and when querying by record identifier, it automatically returns the latest version of the record. This is the most efficient solution for real-time inference.
I measure that the answer is:
Use the SageMaker Feature Store GetRecord API with the record identifier.