Well when you add PDFs and use the get metadata function it tries to extract relevant metadata and figure out the title, author, year and publisher and everything, right? I assume while it tries to read from the contents of the pdf itself, some of this is parsed from the filename (or maybe I'm mistaken). One of the things it does is remove (what it thinks) are the subtitles, and copy just the title to the parent item. This doesn't always work and I frequently have the subtitle copied over instead of the title, which seems to be discarded. There are also multiple cases where the year and journal are in the filename but aren't transferred to the parent metadata. Here's a specific instance, though there are many others:

A file named: "The Road to Eleusis Unveiling the Secret of the Mysteries - 1979 - Journal for the Scientific Study of Religion" gives the parent metadata Title "Unveiling the Secret of the Mysteries," and discards both the year and the Journal. Yet "The Road to Eleusis" is the most important part of that title, not what comes after it. I can only assume zotero thinks that's the subtitle, but it's clearly a little confused. It does, however, copy over the authors which are not in the filename, which makes me think it doesn't even look at the filename.

There are many cases where it removes the subtitle when really it would be great to keep instead. This seems to be often things that come after " - " and ": " and "_", but not always (see above example). But why discard valuable metadata in the first place? In the interest of keeping a short title? I'm personally not quite sure why zotero is set to try and remove the subtitle which seems to be the main issue causing all this fuss in the first place. What's the advantage in that? That is unless, as I'm suspecting, it's not even looking at the filename.

Well there goes any hope I had of not pulling all my hair out. Thanks for letting me know. I think I'll just give up on zotero then.

Yes, in hindsight, it does seem like that's what I was asking for. But I'm realizing this won't work for me because I have a library of roughly ten thousand pdfs I'm trying to organize. Going through them one by one to identify which ones have more data in the filename than in the automated retrieval function is just not something I can afford right now.

I have a related problem - I actually want Zotero to retrieve the metadata and change parent item name for 15000+ PDFs. It had success with many, but there's a large amount of items with nonsensical or even just plain wrong metadata/names. For those, I'd like to revert the parent name to the filename. Is this possible in some way? Perhaps with one of the addons like Zotfile or Zutilo, or some sort of script? It's not really feasible for various reasons for me to remove the files-to-be-renamed and add again

I didn't know that was possible. Anyway, its not helpful here - I restarted Zotero weeks ago.

I'm simply wondering if there's some sort of batch process that can set the parent item name/title to be the filename, while maintaining all the other metadata. Essentially the reverse of "Rename file from parent metadata"

(p.s. I think my comments belong in the original discussion, not this split-off one. Sorry if my description was confusing)

Thanks for the quick reply, as always. I have found a bit of a workaround workflow that will make it possible, but a bit tedious, for me to achieve this.

@jeapordy Unfortunately I don't have the faintest recollection of what I did. Sorry.