A support user starts a baseline upload for an organization. The baseline upload had 100000 rows in the CSV file uploaded

A support user starts a baseline upload for an organization. The baseline upload had 100000 rows in the CSV file uploaded, and the support user - right after clicking the “Start” button - realizes that he seemed to have wrongly exchanged data between two adjacent columns. Subsequently, the support user would need to wait until the current request completes, which can be in the magnitude of hours, and since we don't allow two concurrent baseline uploads in Labs right now, this implies he would not be able to upload the correct file till this task completes. Once done, the user would need to remove those 1 lac entries, and finally upload the correct file.

can anybody explain this to me how can I optimize this scenario