I want to stream big files to the server, avoiding mistakes.
There should be a client and a server.
The client should accept drag-and-drop CSV files. It will read the first few bytes and decide if the headers correspond to any of a number of expected datasets. If yes, it will stream the file to the server. If not, it will complain.
It will receive the file, check its head, and replace the corresponding database table with it. Yes.
Of course.