Would it be possible to store the Zotero files exactly with the same structure as my collection of tabs and/or folders ?

I like Zotero a lot, but i stop using it because i could not find my files so easily or i had to do some tricks to get them out of the random folder structure of Zotero.


The current state is this:
"File Storage

All files stored in the Zotero library are kept in a folder within the Firefox profile directory (unless another location is specified in the Advanced preference tab). Each file has its own subdirectory within the Storage folder, named with a random 8-character string. As such, it is probably inadvisable to attempt to navigate to these files in your file browser. Below this level, though, the files are named intelligibly, so it is possible at least to run a search for them. Easier than all this, however, is accessing these files directly through Zotero. Selecting the item in Zotero and clicking the Show File button in the right column will open your file browser and take you right to the file.
  • Well, your solution doesn't necessarily follow from your problem statement. E.g. there may be other solutions for it.
  • Hi, Thanks. I just made a suggestion. Organizing/searching/archiving for attached files in a random folder structure is not very handy and it would be a great improvement if you could add a different way of organizing attachments in folders.

    Sort of the same feature iTunes offers of organizing your files with a certain order, by artist, etc, all automatically. That's what i would like in Zotero! =)

    I don't use Zotero because it is unpractical to me, but i really like all of its other features and i really used it for quite a while, and recommended it too.

  • Yes, but OTOH, you can quickly access the files from within Zotero (attachments --> show file).
  • You can also drag single attachments directly from Zotero to the desktop or another filesystem folder. (On OS X, you can drag multiple files on once.)
  • edited March 29, 2009
    My solution in Max OS X for the inscrutable file/folder structure of the Zotero directory was to create a smart folder/saved search on my dektop for:

    file type=PDF
    location=Zotero and subdirectories

    Zotero does a nice job renaming PDFs upon import from the journal websites, so I end up with a nice clean self-updating folder of PDFs with "Author - Year - Title" filenames.
  • I agree with Carlos's suggestion that having Zotero arrange files in the same structure as one's collections/sub-collections would be great. Obviously most of the time I access files by opening Firefox and Zotero, but in some cases I need to access files and opening Firefox is problematic (e.g. because it's slow to load and I'm in a rush, or Firefox has crashed).

    amacom73's suggestion is really useful (I'm not sure how to do this in Windows exactly, but there must be a way), but the thing is that with larger collections, having the files arranged in folders is much more manageable - e.g. if I know I'm looking for articles(s) of a specific category, but aren't sure on the author/year/title, searching through a list of hundreds of pdf's is pretty hard.

    Having the files neatly arranged in folders would also be very handy if at some stage I had a more drawn-out problem with Zotero or Firefox, or if I wanted to access my Zotero files on my flash drive on a computer without Zotero and Firefox installed.
  • I'm not crazy about the folder structure either, I don't think, though, Carlos' suggestion is a very good solution:
    Note that collections are _not_ folders, but more akin to tags. I.e. the same item can be in several collections - or in no collection at all. If you remove an item from a collection it continues to exist in the library. etc.

    I currently don't see a great way of dealing with this. My best bet would be that indexed searches on computers (think google desktop search, sherlock, tracker once it works properly) become so fast that the actual storage of a file is increasingly unimportant - especially if it has a meaningful name, which Zotero does by default for files it attaches itself and allows you to do manually for others.

    If this is a big deal for you, you can obviously just store the files outside of Zotero entirely and just link to them, but that is clearly a lot more work.

    A way to possibly address some of the concerns would be to create a smart/virtual folder structure paralleling collections - i.e. just links to items. But even that wouldn't solve the flash drive issue...
  • Ahh, good point Mr Smith, I didn't think of the item-in-multiple-collections issue! I had considered keeping my files stored outside of Zotero, but this would indeed have meant a LOT of setting up individual items + linking, as opposed to just dragging and dropping a bunch of files and grabbing the pdf metadata.

    Well, perhaps the invisible hand comes up with something on this (har har!) Going by a few comments I've read, having easy access to stored files outside of FireFox (in some way) does seem to be a common request.
  • I have made the folder suggestion since i started using Zoter. However, i stop using it mainly because of this reason... For me it is really surprising that they could not improve this feature in more than perhaps two years of development. In such time that feature should be possible to program. Therefore, i may be wrong, but i get the feeling they did not want to improve the folder allocation of Zotero. Too bad because i really stopped recommending Zotero... very simple, there is absolutely no other software i know that does not let you save your files where you want to save them...
  • your tone is uncalled for.

    Obviously Zotero lets _you_ save your files where you want to - but it doesn't allow you to customize exactly where the _program_ saves the files it downloads automatically.

    What alternative do you use? Which program does the same thing better? How exactly?
    As noted above, the solution proposed by you does not work well with the way collections work in Zotero.

    Also - there are a whole bunch of programs who do similar things. Pretty much any e-mail client, for example, saves attachments in weird locations.
    Endnote, for example, does allow you to put pdfs in one single folder and doesn't allow you to specify a folder structure either.
    The one program that does appear to do this better is Papers (based on what I've read - haven't actually tried it)- but it's a one platform solution, which obviously does facilitate issues like that - and it has plenty of other drawbacks.
  • Dear Adam,
    My opinion is that you are not addressing my comment in a good tone and I have the feeling you are not receptive to my comments. This is actually the second reason i stop using and recommending Zotero. I did not feel Zotero developers were listening to my comments and the same comments of other people. I do not know if you are an admin or a Zotero developer and i´m sorry if my tone was somewhat direct.

    Despite all, i am willing to reply to your questions. I´m using BibDesk and i save all my files myself. BibDesk allows me to have the right file format to use with Latex editors and at the same time i can automatically retrieve citations from some websites, not really perfect but OK. With BibDesk i do not need to export my collection before being able to usit with Latex. That saves me one step compared to Zotero. Second, i save my files on my own folders without any software assistance. That is slow but it has the great advantage that i can save them where i want and i just can search for them easily. I can tell that even the Spotlight search of Mac OS has some trouble finding the files in such complicated folder structure of Zotero. Then I can save time when i search for my files which i do quite more often than saving them. I save them once but i search for them quite several times. In this way i save much more time compared to letting Zotero organize my files.

    Indeed, as you say, i do not know any bibliographic manager that is actually superior to Zotero (in this feature!) but that does not mean that Zotero is superior to what i can do myself. On the other hand, as you can notice, people with large collections have this trouble of searching for their files very often and I´m one of these users with a very large collection. If Zotero could save some time for me I would definitely use it but that is not the case at this moment. I´m sure Zotero can be very useful and save time for many other users but unfortunately it is not for me.

    Hope this comment helps. Good luck!
  • I have no association with Zotero. But the claim that the developers aren't receptive is - if you look around on the forum - quite odd. Apart from the large design decision, Zotero seems quite community focused and many concerns that people voice here are taken up and included.

    I agreed above with your statement that Zotero's file structure isn't great, but I also pointed out why the alternative suggested by you is not really straightforward and probably not a good idea.

    As for your solution: if you save your files manually anyway, what's wrong with linking to them from Zotero (which gives you cool features like full text search etc.)? You can turn off pdf downloading for Zotero if you prefer that option. And if your collection is very large, wouldn't it still often be faster to look for a file from within Zotero? (What's the problem with that anyway? You never say - for me that's the main reason this isn't a huge priority: Searching a file within Zotero will always be faster than on the desktop).

    Obviously, for LaTeX users, Zotero is always one step more - though given the far superior scraping capabilities, I would have thought it'd still be a lot faster to export to bibdesk.
  • adamsmith: Mendeley is one example of an app that does quite a good job of this. It allows the user to define a folder structure and naming convention for pdfs, based on extracted metadata.

    I only point this out for interest. Mendeley is a different (and currently much less functional) animal than Zotero, and its task is easier, as it only deals with pdfs (and not all the multiple files associated with web page snapshots, etc).
  • that would be interesting to look at - though once again, like papers, Mendeley just works on one platform (incidentally the one that deals best with pdfs...) so its job is easier.

    Could you give an example, though, of how this would work in practice? Would I have a folder per author?
  • Well, firstly, you can choose whether you want Mendeley to organise your pdfs. If you prefer, it'll leave them where you put them. If you choose for it to organise them, there are options for how you want Mendeley to file them: using author, title, year or whatever. So you could have /author/year/pdf-name or /year/author/pdf-name. Then the pdf file name itself can be customised using extracted metadata.

    Btw they do have linux and OS X builds of the desktop client. I'm not advocating it (I'm mainly a Zotero user), but it's development is interesting, and it seems the team are very open to working with other systems. They have a ticket open for two-way sync with Zotero, and already have a 1-way CiteULike sync option working.
  • yes, they've been here on the forum (Victor, I believe)- seem like great guys and it's a good program. Strongest competition I've seen so far for Zotero.

    I'm not super impressed by that file organizing option, though. That seems only marginally (if at all) better than having all files in one folder. (though I guess that's not the Zotero default either, but as per above it's easilz feasible).
  • Given that they have the folder arrangement happening, though, I imagine it would be easy enough for them to extend it to extend allowing for pdf storage paths to use keywords, for example. Mendeley can currently rearrange the pdfs if you change your folder/naming options, which is quite nice. It also can keep watch on a folder used for incoming pdfs, and automatically move and rename them.

    But, again, Mendeley (at least at present) is a simpler app than Zotero, dealing with fewer item and file types. As you say, it has an easier job.
  • FWIW, some CD burning apps might be a good model for this; for example, IIRC Max. Here you have a preference item that allows you to construct the file path from the data variables and such.
  • adam between quotes " " my replies after >>

    "I agreed above with your statement that Zotero's file structure isn't great"
    >> I would say it is an odd and troublesome structure if you want to do a desktop search. My opinion only.

    "but I also pointed out why the alternative suggested by you is not really straightforward and probably not a good idea."
    >> I do not know what the best solution is.

    "for me that's the main reason this isn't a huge priority: Searching a file within Zotero will always be faster than on the desktop"
    >> This is at least not true in my case as i search using Spotlight in Mac OSX. I need five steps to search within Zotero. 1. Open firefox. 2. Open the Zotero screen. 3. Search. 4. Find the file link of Zotero 5. Click on "show file". On the other hand, if the folders were organized in a prettier way i could use my favorite desktop search software to search for them and at least with Spotlight the files you are looking for are just seconds away to be opened. And i say at least i could use a desktop search engine because browsing through the files outside Zotero is not possible due to the random folder numbering. In summary Zotero does not allow for desktop file browsing and on top of that searching for files easily outside of Zotero is not possible. At least for me these are great limitations of Zotero.

    "Obviously, for LaTeX users, Zotero is always one step more - though given the far superior scraping capabilities"
    >> I agree that Zotero can scrap nicely the publications data BUT again and i repeat that scrapping occurs only ONCE per file, while searching for items happens much more often and several times. In my case i prefer to be able to search faster than to have superior scraping results.
  • And... it would be great to know which "user suggested" features were actually added to Zotero in order to judge whether Zotero developers are really listening to their customers... (and of course, i would not count the bugs reported by users as new features)...

    Good luck!
  • edited July 21, 2009
    but if you're using spotlight anyway, the smart folder option outlined above should work perfectly for you. Plus you can just drag the pdf out of Zotero.

    As for suggested features, of the top of my hat
    - rtf scan and changes to the way it works
    - one click renaming of attached pdfs
    - retrieving metadata
    - adding items by identifier
    - changing the behavior for "create new item" for a pdf
    - rich text notes
    - duplicate detection
    - a lot of the work on the new csl processor - while outside of zotero - has been driven by community input

    those just a couple of examples that have come up recently - probably easy to think of many more.
    As for scraping vs. searching: I don't see how your current manual technique could not be used with Zotero? As I said - just turn off automatic pdf and snapshot download!

    @Bruce and CB - OK that sounds like it could be feasible - does the dev team have any input on how hard something along those lines would be?
  • @amacom73 - I'm trying out the saved search method you kindly suggested, but the pdf's aren't named nicely with Author/Title/Year :s The ones saved directly in Zotero (as opposed to my pre-Zotero collection) are just named with the filename the database I've downloaded them from chose to bestow (e.g. "science.pdf", "fulltext.pdf", "4731/80579ba" etc). Do I need to be saving pdf's in a particular way (e.g. "new item from current page" vs taking a snapshot etc) to be getting the file named with Author/Title/Year?
  • yes.
    Zotero only automatically names the item if you use the translator:
    i.e. use the icon on the right of the address bar.
    For other icons, after you have entered the citation information, you will need to right-click and then choose "Rename from parent Metadata"
  • Brilliant, thanks :)
  • Would it be possible to create an export feature in Zotero in which pdfs and other file types are output to a master directory organized in the same structure as Zotero collections? Items that appear in more than one collection could just have a "shortcut" file written in place of the subsequent "duplicates" in the folder structure. This would allow users a way of keeping a copy of their collections on a USB drive and not have to use Zotero to find/navigate to teh correct file. It would be great if such an export function could apply to the currently selected collection and any subcollections. This would enable users to share a subset of their collections with colleagues who do not use Zotero.

    The reason I see this as important from a user perspective is that I already have a large collection of pdfs that I've put into folders. I'm trying to import this into Zotero. To avoid having duplictes of files on my hard drive, I delete the "originals" and rely on teh stored version in the Zotero storage folder. The export feature would offer peace of mind to users should Zotero no longer be supported in the future for some reason. This would allow me to eventually go back to my old (and less functional) system.

    One of the great strengths of Zotero is that it is an open system. Having an export option that allows users to easily move to other systems or plan for software obsolescence in the long run would further strengthen Zotero's openness and lower the barriers to user adoption.
  • Carlos , i have the same feeling as you regard the "folder" thing.

    We all have an Operation System(window,osx,linux...), and we get used to the way we manage files under them. We want to use this way as long as we could and special software should help us achieve maximal efficiency of that system not try to replace the OS. In the field of PIM, they have the same problem. They found that it is maybe one of the biggest reason people abandon PIM software. Here is an example of the direction i said.

    I think your suggestion is brilliant but even better if the software can do this as soon as i change something in Zotero.
  • > My solution in Max OS X for the inscrutable file/folder structure of the Zotero directory > was to create a smart folder/saved search on my dektop for:
    > file type=PDF
    > location=Zotero and subdirectories

    First of all I agree with all who are asking for a better folder structure for the document storage in Zotero - for the same reasons (I need access to the files from without Firefox/Zotero). I have a big folder with all my pdf's and started importing them into Zotero. Now I realized that I have a hard time knowing which pdf's are already in the Zotero database, especially when I add new pdfs to this folder.
    I tried what was suggested above but for some strange reason Spotlight does not like the folder structure of zotero/storage and returns zero files. I tried all options, confirmed that the files are there... nothing! Someone knows what's going on here?
    I have the latest version of snow leopard installed.
  • In WIndows, I use Everything (that's a program name, from -- a quite brilliant program that indexes filenames incredibly fast with very low cpu usage. You can easily create a saved search there for the zotero storage directory, which solves many of the problems discussed above. You can't pick out particular collections, of course, but since items can be so easily moved among collections and duplicated in multiple ones, it would be very hard to maintain synchronization between the zotero data and a folder structure no matter how you did it.
  • Just adding to "petition" -- I too would prefer that PDFs be arranged in a more user-accessible folder structure (a la iTunes, or Papers) for reasons already well-outlined above. I just started using Zotero and the opaque folder structure was the first thing I noticed.

    Also, since many users have existing PDFs arranged in one's own sensible folder structure, it would be good if Zotero let you import PDFs without copying the actual file to Zotero's database (a la iTunes' option to "Let/Prevent iTunes from managing my music files"). This would both allow one to maintain one's existing file organization, and prevent the duplication of PDFs when one imports existing PDFs into Zotero.

    Lastly, I appreciate the sensible naming of PDF files. However, I often do not remember a paper by its first author (typically a name I'm not familiar with), but rather by its last author (typically the director of the lab). It would be good if Zotero had an option to name files with the names of all authors (when fewer than some number N) or perhaps with the names of the first and last authors. Thoughts?

    Thanks for an otherwise terrific product and service.
  • the reason Zotero has the files in its own folder structure by default is that this is the only way file storage/syncing works and can work.
    But you can already link files, although currently linking pdfs by drag&drop doesn't work yet, but is frequently requested and planned.

    As for the folder structure - in general yes, but as per above, no one has been able to convince me why using a virtual folder doesn't solve 99% of the problems.
Sign In or Register to comment.