If the csv format is the same why create new tables? Having a staging table which is cleared at the start of the whole import and just adding another column or two to indicate in some way which csv file that row came from and the import date/time would be more effective surely?