You can't have two files with same name in same folder, so I don't understand your doubt here!!
If you later move the files to the central repository, files names must be unique, or you are going to overwrite the ones already there.
In the PDFE database, metadata is linked to the full path+filename, this full path+filename is what's used as uniqueness identifier, exactly what happens with the file system.
Or you have not understood my last post, or I'm missing what you really want to accomplish, and/or the method you what to use.
What's usually done with PDFE is to have all the files in a central files server. LAN PCs can concurrently edit these PDFs metadata across the network, maintaining a local database, but because metadata is also committed to the PDF files itself, the main PDF Explorer instance, the one serving the web interface, only has to maintain the index actualized.