DEA-C01 Practice Test 8

An Amazon Data Firehose stream delivers JSON events that contain a customer_id field to Amazon S3. The company has about 300 customers. Analysts query the data with Amazon Athena and almost always filter on a single customer, but every query scans the whole dataset because all objects land under one prefix. The team wants the data grouped into per-customer prefixes as it arrives, without operating a separate repartitioning job. What should they do?