You can use the "Last Update Time to start". When you check this checkbox it will extract the data which has been created or updated since last time the extract rule was run. This way it will not extract whole data every time and will not have repeated data.
Another option you have is you can use an sftp location(repository) to store the extracted data instead of storing it to the server.
Another option is you can run a agent at some interval which fits your business requirement to cleanup the folder where you are storing the extracted data on the server.