MLA-C01 Practice Question 62: Firehose Lambda Transform
AWS Certified Machine Learning Engineer – Associate (MLA-C01) · Test 7 · question 2 of 10
A logistics company sends JSON telemetry from delivery vans through Amazon Data Firehose into Amazon S3, where it is later used to train an ETA model. About 3% of events are malformed and should be discarded, temperatures arrive in Fahrenheit but must be stored in Celsius, and a nested gps object must be flattened into top-level fields. The cleaned records must land in S3 within a few minutes of arrival, and the team does not want to run any servers or clusters.
What should the team do?
Show answer and explanation
Answer:
Firehose data transformation buffers incoming records and invokes a Lambda function on each batch before delivery. The function returns every record with a result of Ok (transformed), Dropped (intentionally discarded), or ProcessingFailed. Firehose treats Ok and Dropped as successful and delivers only the transformed records. Unit conversion, flattening, and filtering happen in the stream, the cleaned data lands in S3 within the buffering interval, and nothing needs a server (option D).
- A is wrong because the malformed, Fahrenheit, nested records are what land in S3. The cleaning happens again on every query, and the stored training data is never clean. - B is wrong because an EMR cluster is exactly the infrastructure the team wants to avoid, and it replaces a managed delivery stream that already works. - C is wrong because S3 Object Lambda transforms data only when it is *retrieved* through the access point. The objects in the bucket stay dirty, and every reader must know to use the access point.