I\'m not sure how it works so I\'m looking for the right solution. I think SSIS is the right way to go but I have never used it before
Every mornin
SSIS Sounds like the way to go. The way I've seen your type of issue handled previously is with a Staging Table. The new document loads into the Staging Table- then Staging and Production are compared- obsolete records are archived (not JUST deleted) from Production, existing rows with some changes are updated (again, original data archived somewhere), and new rows are inserted.
Note: Your definition of "obsolete" needs to be very, very precise. For instance: should something be archived away just because a matching row does not exist in your most recent file? Should it stay for X amount of time in case it comes on a subsequent file? These and other questions should be considered.
Almost any standard SSIS tutorial should point you down the correct path for how to do each of these steps.