problem with duplicate entries

edited October 26, 2017
hello. in zotero (last stable) i have item A & item B created from scratch and not by duplication process (i'm 99.9% sure they wasn't created by duplication).
Zotero tells me they are duplicate
So i create item C, and copy by hand item A content to item C.
Then i create item D and copy by hand item B content to item D.

then i move to trash and definitively delete item A e B.
result is: zotero tells me that Item C e item D are duplicates.

items are book chapters with same same title, same first author, same date, same publication place, same press, but different editors, different pages, different book title and different isbn.

How to resolve the problem?


  • There isn’t a way to tell Zotero that these items are different (it’s getting caught up on the same title, Author, and year). A feature to mark items as non-duplicates for such cases is planned.
  • It's not really a huge problem: Zotero's "Duplicate" function is merely a way to help you find and identify duplicates. As long as you don't merge them, Zotero still treats them as completely separate items. But yes, the function does occasionally produce false positives and at this time you cannot mark them as non-duplicate
  • edited October 26, 2017
    it's a problem when you have a large library...
    it could interesting to implement a function in the right click menu to "separate duplicated items" that generate another ID to the one created from a duplication process. often i need to duplicate an item for made a little modify in a fast way then separate the items to have 2 items with 2 different IDs but little different content. the same "separate" function could be useful for resolve false positive too.
  • the moment you duplicate an item, the two have separate IDs and items created by duplication are no more likely to show up as duplicates than items that are similar for another reason. This really boils down to not having a way to mark two similar but not duplicate items as non-duplicates. That doesn't currently exist, but you're not the first person to ask for it. This is obviously technically doable, but it's not trivial and I'm not aware of any near term plans to implement it.
  • I think the duplication finder algorithm and/or the merge algorithm are bugged.

    Items A & B (book chapters) differs in:

    book title
    (item creation timestamp)

    but the merge function allow me to choice the right version only for book title, pages and isbn fields: for the editors field it doesn't allow me to choice between the two differents editors fields.

    it's a known behaviour?

  • That's correct -- you can't merge creator fields. I believe because it's too complicated GUI-wise to get that right
  • edited October 26, 2017
    Would be comfortable a notification message about it, that inform the user about the merge exception for creator fields, because in the hurry you can go wrong believing that creator fields are the same when they are maybe slight different.
    In my opinion, certain zotero functions (like merging or import/export) makes you feel a security that is actually missing, and a notification message alerting you about possible inconsistent or bad behaviours would be very useful.
    I know, for example, that the zotero manual explains about import/export limitations & strange behaviours but not everyone read that specific manual part, because the expected behaviour of an import/export is data consistency and the last thing an user has it's about a possible data consistency problem of a "primary function" of a stable program that reached the 5 version.

  • @adamsmith Would it be possible (as short-term easy fix) to have the duplicate function:
    - look in the Extra field and if it finds a "Duplicate: NO" AND
    - the items identified as duplicates are tied by the "related" feature THEN
    override the duplicate identification?
  • I doubt it. At that point it's not much more work to implement this properly. The use of the extra field is currently very consistently just to map CSL variables. There's limits of how far we'd want to push that.
  • OK. It did not sound very elegant even while I was typing it out :)
Sign In or Register to comment.