The existing dc identifier should be unique. The primary one is used in font deobfuscation/obfuscation. It can be used in the ncx as well and must match the opf one if an ncx exists.
A random uuid primary dc:identifier can be generated for new empty epubs.
But using the template epub plugin would be even easier as it was designed to handle that.
https://www.mobileread.com/forums/sh...d.php?t=341735