I want to create an S3 connection that will have monthly data stored in individual csv files that is connected to ADO. Around month end, data will be uploaded multiple times, replacing existing csvs. When this happens I want to replace the existing data for that month with the new file. I don't want to have to create a primary key for every dataset in order to do an incremental load but wanted to check if doing a replace will reload all of the data for every month or just the files that have been uploaded?
In addition, if we do have to create a primary key for the data, can it be a combination of columns or 1 single column
