Choosing "Reload only" will use previously staged data if it's available. Note that reloading without schema regeneration might lead to missing columns or tables.
This option loads the most recent initial load data, either from when the dataset was first added to the pipeline, or from when the last "Restage and reload" was completed. This load does not include any delta loads that have occurred since the last initial load. This action is an attempt to re-write it to the database again, using the schema we previously inferred. If there is no previous data to reuse, then Precog will stage and scheme the data anyway.
Why - useful when configuring the pipeline for the first time. A customer has changed something about the destination configuration. They want to reload the data into the destination but do not want to wait to pull the data from the API again.