Choosing "Reuse schema" will reuse previously generated schemas for selected datasets.
This option is only available after the first data load has been successfully staged but not necessarily written to a destination database. This action will pull the data from the API again from scratch. The original schema will be maintained. Then Precog will attempt to rewrite the data to the database again.
Why - a customer has changed a start date in the connector (source) or there is some other problem in the data, that would necessitate pulling the data from scratch again. The customer is emphatic about what the schema of the table looks like (name and structure or just name?) in the destination and does not want this to change under any circumstances. Many times this is because there are downstream processes or reports that rely on the schema names. If those names change then everything downstream stops.