Mendeley Import!

zeynep
I'm a long time Zotero users and my Twitter feed is overflowing with people who want to switch to Zotero from Mendeley. First problem is they think of Zotero as it was years ago--so Zotero needs a Mendeley/Zotero comparison page.

However, worst issue is how to get Mendeley data into Zotero. I understand some of the problem stems from the Mendeley side but it would be really really great if there was a tool that automated as much of the process as possible and a guide to how to do this. I think it is important for Zotero to step up this moment and help people come over!
«1
  • adamsmith
    edited April 9, 2013
    Hi Zeynep,

    Mendeley Import:
    This is really straightforward:
    1. In Mendeley go to all documents, select all, then under files select "Export" and export into bibtex (.bib) format.

    1a) Windows users only: Due to a bug in Mendeley's bibtex export you will need to follow my instructions here to correct filepaths in Mendeley's bibtex exports

    2. In Zotero, click on the gears menu, select import and select the exported Mendeley .bib file. Import for large libraries may take some time. In rare cases you may get a busy script warning, click continue.

    There is a guide with pics here: http://ica.library.oregonstate.edu/tutorials/lesson/631--Advanced-Zotero?mid=26031&type=MiscellaneousResource&uid=1271 the author uses RIS instead of bibtex, which saves Windows users step 1a) but may give you slightly worse data. If you have mainly books and articles the difference likely won't matter.

    What this will _not_ import is Folders. The only way to do that is to export/import them individually into Zotero. Since Mendeley doesn't export folders, the only way to import them would be to go through the sqlite directly and that'd require a very significant amount of coding.


    Comparison Tables
    I find comparison tables made by software companies dishonest - I know they all do it, but I think tables that just show how your product can do everything that other products can and so much are basically a lie (because obviously Mendeley has features that Zotero doesn't and vice versa, same for Endnote, Refworks etc.). I don't think Zotero should join that unfortunate trend.
    This by Cambridge:
    http://www.jbs.cam.ac.uk/infolib/2012/10/29/zotero-versus-mendeley-2/
    or this by UW Madison
    http://www.library.wisc.edu/citation-managers/comparison.html
    are decent, though those lists outdate fast (see e.g. Aurimas's comment below the Cambridge one and the fact that the UWM says Zotero doesn't support abbreviation lists, which it does since 4.0.
  • zeynep
    Any chance for developing that code to automate the export folder? I know it's a big investment but sounds like it would be really, really helpful to many people.
  • adamsmith
    In general yes - Aurimas and I are both interested in extending import to include direct database import from Mendeley and Citavi - but this isn't going to be quick, even though it'd be nice to have it now when all the people are looking to switch.
  • zeynep
    Define not quick? A week or month or three months?
  • adamsmith
    definitely not week or weeks. Likely multiple months.
    Zotero is a small project - things like that done mostly by volunteers (like Aurimas and me) on the side of our day jobs.
  • adamsmith
    See also the suggestion for dealing with long filenames here:
    https://forums.zotero.org/discussion/28808/importing-for-former-mendeley-users/#Item_0
  • dstillman
    One potential method for dealing with folders would be to assign tags to items based on folders in Mendeley, export, and then recreate the collections in Zotero based on the tags. Still annoying, but at least a relatively straightforward way to preserve the folder structure (at least for people without a huge number of collections).

    Local Mendeley SQLite databases are essentially unusable, with a DB schema that looks like this and data with all sorts of proprietary junk (e.g., "<m:linebreak/>" for newlines in notes), so I don't think trying to use that directly is worth anyone's time (which, being charitable, maybe was the point).

    Their BibTeX and RIS output doesn't include unique identifiers, so it's not possible to post-process the output in combination with the database to get items' folders, but an approximate version of that could be done with titles and creators.
  • adamsmith
    FWIW, Mendeley suggests doing this via API
    https://twitter.com/mrgunn/status/321666531295637505
    Colwiz apparently does that, though I'm not sure how well it works.
  • aurimas
    I'm looking into writing a zotero plugin for transferring mendeley library via their api. probably won't get to it until the weekend though.
  • DWL-SDCA
    This seems to have become more important due to the reaction after Elsevier's purchase of Mendeley.
  • adamsmith
    Yes. See Zeynep's post at the top that prompted this.
  • aurimas
    Comparison Tables
    I find comparison tables made by software companies dishonest - I know they all do it, but I think tables that just show how your product can do everything that other products can and so much are basically a lie (because obviously Mendeley has features that Zotero doesn't and vice versa, same for Endnote, Refworks etc.). I don't think Zotero should join that unfortunate trend.
    This by Cambridge:
    http://www.jbs.cam.ac.uk/infolib/2012/10/29/zotero-versus-mendeley-2/
    or this by UW Madison
    http://www.library.wisc.edu/citation-managers/comparison.html
    are decent, though those lists outdate fast (see e.g. Aurimas's comment below the Cambridge one and the fact that the UWM says Zotero doesn't support abbreviation lists, which it does since 4.0.
    I realize that it's hard to stay objective, but I think it would actually benefit everyone to have an actively-maintained up-to-date table listing features from top reference management software. Just having to maintain the table would give _us_ a better feel for where we're lagging behind.

    Again, this would have to be kept objective, so we may need to come up with a list of criteria to test and a proper way of doing it. E.g. I would definitely like to know if Mendeley or Papers outperform Zotero in PDF metadata retrieval. I don't think they do, but we should have an objective way of testing this.

    All of this should go into a separate thread though.
  • ljvillanueva.coquipr
    This won't be automated, but I managed to move my 1600+ reference library without much problem. My experience is here: http://research.coquipr.com/archives/492

    If you have several folders, I would recommend doing it one folder at a time.

    This may not be ideal, since it can take some time to get organized and re-create the folders in Zotero, but it works.
  • c.rusbridge
    Oh dear... I started moving stuff from Mendeley to Zotero in smallish chunks (to give myself confidence, and to try to understand how things work), and also moved a large chunk of stuff twice (once via bibtex and once via RIS). In the latter case I _thought_ I had deleted both the first collection and its items before doing it again. Now I seem to have most of my items as duplicates, but often of different types. I was going to go on and select the whole (1200 item) Mendeley library, but that's going to give me yet more duplicates of the 50 or so items I've already imported.

    Is there a way to wipe the Zotero database and start again?
  • aurimas
    You can delete the imported items, empty the trash, and import again. Make sure you are actually removing the items and not just the collection (collection is more like a tag than a folder in Zotero). If you click on My Library it will show you all items in your library.
  • dstillman
    Actually it's better to just delete zotero.sqlite and 'storage' in your Zotero data directory in that case, since then there won't be thousands of entries in the delete log that could slow down syncing. You can just delete those files and restart Zotero/Firefox and you'll have an empty library.
  • schmidli
    Hi! Tangentially related: I'm a librarian at UW-Madison. Our comparison chart has become quite unwieldy, and we definitely fall behind sometimes. So, we welcome any suggestions for ways to make our comparison chart better and more accurate. You can contact me directly with suggestions, or contact our group here: http://library.wisc.edu/citation-managers/get-help.html#submit-a-question

    I'll update our mistake re: abbreviation lists! :)
  • adamsmith
    (to prevent duplicate efforts: I've sent schmidli a longer note with comments on the chart).
  • s.emrys
    Hi

    I'm still trying to escape mendeley's clutches, but put off by the hassle of individual folder export-import. Was there any progress made on an automated process / plug-in for this?
  • adamsmith
    nothing yet, no.
  • s.emrys
    And is there any timeline for a likely release?
  • aurimas
    Sorry, no timeline, but I'll try to get to it soon (within a month or two). Some other things I'm working on currently.
  • s.emrys
    that was quick! OK - glad to hear its on the timeline. Thanks
  • s.emrys
    Hi Aurimas

    Just checking in, as I heading into a writing phase - what's the timeline looking like for a plug-in?

    Cheers
    Emrys
  • TTLP
    Just to ensure centralisation of topic, here is the following link:

    https://forums.zotero.org/discussion/30349/import-pdfs-from-mendeley-to-zotero/#Item_19
  • cl_cam
    Hi adamsmith,
    there's an updated version of the comparison table Zotero vs Mendeley by Cambridge - only published this week:
    http://www.blogs.jbs.cam.ac.uk/infolib/2013/11/12/zotero-versus-mendeley-3/
  • adamsmith
    edited November 14, 2013
    see aurimas below for comments.
  • aurimas
    edited November 14, 2013
    Couple notes:
    • Zotero also works with Opera now

    • It's a plugin, but ZotFile can extract annotations made in most popular PDF viewers = you choose where you want to annotate your PDFs

    • Zotero's mobile site is quite good, so you can pretty much use it as a mobile client for browsing/editing metadata and accessing file attachments (if they are stored on zotero.org)

    • "Can automatically re-name and organise PDFs". Again, ZotFile can do this for Zotero as well
    • Exporting directly from databases (save as you browse) is a bit misleading since that list is far far far from complete and is missing some big players like EBSCOhost and most of the life science journal websites. So maybe include an ellipsis or "Many more..."
    Edit (adamsmith's suggestions):
    • http://www.papershipapp.com/ is missing for both Zotero and Mendeley

    • Under Access it says "Online:" for Zotero. All of the options listed actually deal with locally stored database. Although connectors _can_ save directly to the online database, that is not their primary function.
  • purwoko
    is export from mendeley to zotero including the file?
  • adamsmith
    attached files? yes.
This discussion has been closed.