Sync Error: Long author list for countless papers

patel93
I have an issue with long author lists for many papers in multiple Zotero folders. Manually sorting through the author fields would take a great deal of time.

Does anyone know a way to automate/circumvent the syncing problem of papers that have many authors?

https://s3.amazonaws.com/zotero.org/images/forums/u5604797/yjdzspxguaarhh4p5j46.png

I realize this is an old error to mention on this forum, but it seems so important that I can't believe it hasn't been correct thus far.
https://forums.zotero.org/discussion/comment/280209#Comment_280209
  • dstillman
    To be clear, this isn't a bug — it's a data import problem, usually caused by bad data exported from some website or tool. The desktop app allows you to import the data and fix it later, but the online library needs to have some limit on the length of fields.

    Where did you import this data from, and how exactly did you import it? E.g., saving that item from PubMed via the Zotero Connector imports the authors properly.
  • adamsmith
    (Almost certainly an OVID export -- you'll recall this keeps happening and we ought to try to detect it on RIS import.I think)
  • dstillman
    edited 5 days ago
    (Yeah, but I thought OVID (and maybe just PsycInfo via OVID?) was exporting invalid editor data, and we were going to just truncate it. These are valid authors from the paper that we wouldn't actually want to discard.)
  • patel93
    I almost always use the Zotero Connector browser extension from a Chrome-based browser or Firefox to import PDFs into Zotero.

    I've been doing so for years, but lately the multi-author papers are more common in my discipline.

    For now, I'll manually sort through the poorly imported documents and maybe code up an extension to flag the bad ones. If there was a way to alert users during and after import (immediately), that would help.

    I don't use OVID much these days and occasionally use PubMed.

    More often, I download preprints and straight from journal article webpages.
  • dstillman
    But I’m asking where exactly you saved the item from. Again, this doesn’t happen from the PubMed page, and it should never happen when saving from the Zotero Connector. If it does, we can fix it. The Zotero Connector fixes all sorts of bad data served by sites.

    What are the values of the URL and Library Catalog fields?
  • patel93
    Ah, well there are many troublesome articles in my Zotero account now.

    On MacOS, using Zotero 7:

    Article 1
    URL: https://arxiv.org/abs/2406.16253
    Library Catalog: missing, no text shown here

    https://s3.amazonaws.com/zotero.org/images/forums/u5604797/6cqfeop5bfibdg5fzd3f.png
  • dstillman
    Right, so you imported all of those from a file exported from somewhere. Those weren't saved from the Zotero Connector.
  • dstillman
    edited yesterday at 4:22am
    You should figure out what tool created this data and let us know, but our options to fix bad data like this automatically aren't great.

    Zotero does have a tool that lets you split or delete long tags that were incorrectly concatenated in exported data, and it automatically shows that tool on sync when necessary. We should probably add a similar feature for creators, and extend it to run on multiple items. That's probably the best we can do.

    For your current situation, I've provided a script for people that just removes long creator entries, but that was in the case of Ovid where, as I say above, it was just junk data. These look like valid authors, so you wouldn't want to do that. If you have a lot of these, someone (or you) could adjust that script to convert these to separate creators. Alternatively, if you haven't done anything with these items yet, the easiest option would be to just sort by Date Added, delete the entire batch of items with invalid creators, fix the file you imported with Find/Replace, and reimport.
  • patel93
    "Right, so you imported all of those from a file exported from somewhere. Those weren't saved from the Zotero Connector."

    I am not sure what this means. If I click the Zotero Connector (Chrome extension) in my browser and find that the metadata and PDF are copied over to my Zotero account, is that not using the Zotero Connector?

    For my fix, I'll play with scripting a special solution using some kind of LLM.
  • adamsmith
    If I click the Zotero Connector (Chrome extension) in my browser and find that the metadata and PDF are copied over to my Zotero account, is that not using the Zotero Connector?
    It is, but it is virtually certain that that isn't how this item was imported into your library. You can test this yourself: if you click on the connector when looking at the above arXiv URL, you get all 40 authors listed individually in the metadata and arXiv.org in the Library Catalog field. The Library Catalog is almost always populated when using the Zotero connector.

    You get the same if you get the PDF directly from your browser (that's not always the case, but it is here).

    If I had to guess, this looks like bad RIS from somewhere. The arXiv (Cornell University) is particularly odd -- that's not how arXiv brands itself.
  • aborel
    Note that the authors are perhaps not the only problem with these records: the item type for the selected reference in the screenshot should be Preprint instead of Journal Article.
Sign In or Register to comment.