[Comfy.n]

I did a quick experiment today, as I wanted to get some insight on the relevance of ISBN identifier.

Selected a azw3 file in my library, copied it out of library folder, then dragged it into Calibre as a duplicate. Stripped all metadata from it, including author and title, replacing those by "X". Then ran the ISBN extract plugin on it. The only metadata value this book had, this time, was the ISBN.

Then I ran the meta download from GR, and it worked, getting all the details correctly (title, author, amazon id, comments, grrating and grrvotes). So I guess it's indeed worth running the Extract ISBN first...