So yes, RDT can do this. Depending on your operating system, you may encounter a file lockout issue where UE can’t read the file because another stream has it open already. In this case you would need to create a new file each time but isn’t necessarily an issue that will affect you. It’s a Windows thing. And as long as your python script closes the filestream after it updates the CSV you shouldn’t have any issue at all.
RDT will read the whole file and close it each time then add all results to an array, so to get the latest one in your application you just get the last element of the array. Each element corresponds to one row of the CSV.