Amazon web services – Naive DynamoDb question: How does dynamodb look data up within a partition for a get request?

I know that when executing a get request, dynamodb would leverage the partition key to determine what partition the row is in. But, my question: what happens next? The docs say that if you have specified a sort key, then data within the partition would be sorted which means quick lookup. But let’s say if I haven’t specified a sort key. What happens next? I’d imagine it’s either one of the following but not sure which one is it:

1. there’s a separate lookup (ie O(1) operation) happening (again using partitionKey) within the partition. I’m not sure what hash function would be used here and how it would (or even could) be any different than the previous hash function.

2. DynamoDb scans through the data within the partition until it finds the row matching partitionKey (which seems painful). I’d pray that data would at least be sorted within partition by partition key else this seems like O(n) operation to me. A scan would conflict with popular opinion that DynamoDB is fast for all use cases (with single digit milliseconds latency).

I’m not seeing any docs around this specific case so would love to hear from DynamoDB experts here. Links would be appreciated.

Thanks in advance