Importing PubMed txt; number of items imported incorrect

Hi devs, getting one of two failed outcomes when attempting to import PubMed results that were exported as a .txt.

The first is a successful import that is missing most items. The Import screen reports that 441 items were imported, but what is imported is a tiny fraction. Importing the file into two libraries the first import netted 24 items, the other 15.

Screenshot of Import dialogue and 25 imported items:

Repeated the process, only 15 this time:

The other outcome is import errors using this file, such as 1581327832.

PubMed txt:
  • (FWIW, importing as tab-delimited using EndNote works.)
  • It looks like those items may have eventually imported as there is a tonne of items in the "Unfiled items" view, which should be empty as I always import into a new collection.
  • So if I understand correctly, the only remaining issue is that the import went into the library root rather than the specified collection, yes?

    If you can reproduce that, we can troubleshoot, but if that happened just once inexplicably, it's probably not worth (nor possible) trying to figure out why.
  • edited May 27, 2022
    I deleted the collection and its items when I figured that Zotero had not imported all of the items correctly, which is why I think they ended up in the library root. I don't know what was causing the import to be so laggy as I tested it again now and it imported all items immediately.

    I'll give the .txt import another go later but this might be one of those inexplicable once off things.
  • Exporting from PubMed to an nbib file then importing into Zotero works flawlessly.

    Select the PubMed records you want
    Send to : Citation manager
    Create file

    Use the Zotero file Import by selecting the nbib file you downloaded.
  • I typically export to nbib, but I did not generate this file.

    This may have just been a once-off issue, but the importer shouldn't accept files it can't handle. The import is already slow as hell compared to EndNote, it shouldn't be flaky as well.
  • the importer shouldn't accept files it can't handle
    I'm not sure what you mean by that. It doesn't, obviously. As you saw, the file imports perfectly fine in general. (And for me it imports 441 items into an empty library in 16 seconds and into a 8K-top-level-item library in 20 seconds, which doesn't exactly seem "slow as hell", though we do plan to improve import performance.)
    [JavaScript Error: "operation timed out"]
    Something was happening during your import attempts that resulted in a timeout. Hard to say what at this point, but if this is happening regularly you can generate a Debug ID for an import and we can take a look.
  • Took mine 26 seconds to import into an empty Group Library that I use to remove duplicates. For reference, EndNote imports the same in less than a second.

    Importing is not a dealbreaker because it is not something I do frequently. Good to hear it is about to get faster but it won't be what drives me back to EndNote

    At the risk of derailing the discussion, the biggest sticking point for me is the citation integration with Word. Using the Add/Edit citation button takes ~5 seconds before any results appear, and this adds up quickly in large documents. Again for reference, this is instant in EndNote. I've looked in the documentation and Knowledge Base and there is no info on how to speed this up. I've been meaning to make a new top-level thread about this in case it is useful for others but is there anything that I can do to make the wait shorter?

    > if this is happening regularly you can generate a Debug ID for an import and we can take a look

    Will do. It did happen again just now when I tried to import but I hadn't enabled logging. I'll give it a beat and will try again with logging on in case it repeats. I've had three or four of these JavaScript errors today trying to import the same file.
  • At the risk of derailing the discussion, the biggest sticking point for me is the citation integration with Word.
    You should start a new thread for this with a Debug ID for an operation that's slow. In a large document, make sure you have "Automatically update citations" disabled in the plugin's Document Preferences window.
Sign In or Register to comment.