Reprocessing all records - DynamoDB Stream

Question

I am using DynamoDB stream with a Aws lambda function + firehose to sync my data with redshift. I would like to know if it's possible to add all DynamoDB records to stream to reprocessing purposes. If not, what's the right approach?

Answer 1

For new data,you can do this. For history data,you'd better don't do this. You can dump you table first, then import.

Answer 2

For reprocessing old data a parallelized full table scan is the way to go. There is the matter of deciding how to handle the transition from "old data" to "new data" but that could be achieved using either a timestamp attribute if one is available or by stopping writes to the table if that is possible.