First thing I did after getting the csv export was throw together a python script to save all the URLs as PDFs based on the title. Not horribly robust, but a csv list of failed URLs will be ready when I wake up in the morning.