There is "update:insert if new" option for the output data tool if using an ODBC connection to write to Redshift.
This option really needs to be added to the "amazon redshift bulk loader" method of the output data tool, and the write in db tool.
Without it means you are forced to use the "Delete and append" output which is a pain because then you need to keep reinserting data that you already have, slowing down the process.
Using the ODBC connection option of the output data tool to write to Redshift is not an option as it is too slow. Trying to write 200MB of data, the workflow runs for 20 minutes without any data reaching the destination table. End up just stopping the workflow.