Feature request: Update or force recheck of metadata

01baftb

I switched from Mendeley, and one handy feature they have is for a given item, I can force recheck the metadata from the DOI/PMID.

188BET靠谱吗Would it be possible to implement the same feature in Zotero to force recheck or update metadata of an item from DOI, PMID, etc?

188BET靠谱吗The current work around I have found in Zotero is using the "Add Items(s) by Identifiers" feature to create a new item using DOI.Then I have merge the newly created item with updated metadata with an existing item with missing or out of date metadata.

dstillman

Yes, this is a common request and will definitely happen, though we don't have an ETA at the moment.

ronaldolobato

Great!This is a good thing.

phil.renaud

I agree, this would be a great feature.I am looking very much forward to seeing it implemented.

gagarine

This would be a life saver.

bwiernik

188BET靠谱吗For now, the fastest solution would be to go to the publisher website (e.g., using the CrossRef lookup button in the Locate menu [green arrow in upper right corner in Zotero]), import the item again, then merge the two items.

Another option is to automatically look up the DOI using the DOI Manager plugin, then copy the DOI and use the Magic Wand tool to add an item using the DOI, then merge the two items.

helderc

yes would be great to have this feature.

bjohas

+1 from me!

delogloscj

I'm going to bump this.When proofing 100+ reference documents this is a must.188BET靠谱吗At present, my team uses Mendeley for document writing because of this reason but if this feature were available it would be Zotero all-the-way, premium account and all!dstillman, any update on the ETA?

whatups

Another +1 here.188BET靠谱吗This is the one thing that's preventing me from going all-in with Zotero.Collaborators send me PDFs especially literature dumps to kickoff new projects.Some have metadata, some without, some with garbled metadata.188BET靠谱吗Zotero cannot always get the metadata from the PDF so being able to force update/check the Zotero entry already created with a manually provided DOI would be nice.

Currently using the solution mentioned by bwiernik, which is not too bad when dealing with one-offs.The problem with that is trying to select the two entries (one created by the imported PDF, and one created by either the Magic Wand DOI or from importing the item from the publisher website) so that you can merge them.They may or may not appear as duplicates depending on what has been extracted from the original PDF.And you'll have to hunt down both entries before selecting and merging them.This is a particularly painful for literature dumps.

seeingtheforest

I'll add my support to this.I have thousands of poorly formatted pdfs to import and i have isbns for many of them.

188BET靠谱吗For now I will probably MacGyver something with import to calibre, download Metadata, export to a csv somehow, manipulate it all and create a Bibtex file with everything to import into Zotero.

richardmcgee

This is a must have feature.Any eta?

a-gn

Adding support to this, it would be very useful to refresh metadata per batch.

olilanin

This is a must have feature for me.For one, im in a very similar position as @whatups, my colleagues and collaborators work mostly with literature dumps, and sadly not all pdfs are OCRs based.

188BET靠谱吗I moved to zotero by importing my old bibtex files, and it worked great but so many of the old entries did not have DOI numbers or full metadata.It would be nice if a can batch und just update those files.

RonnenLevinson

I agree that this would be highly valuable.In generating a report we are about to deal with about 1000 references collected by 20 authors!a batch refresh of their metadata would be prudent.

whatups

@dstillmanany update on this feature?

ckemere

Big +1 from me.I began migrating from Mendeley and then discovered that my standard workflow of using PMIDs to update records that were incomplete, i.e., from pre-pubs (and also to pull in abstracts), didn't work.The normal process is that I read a paper when it first comes out, then later want to edit a record in order to cite the paper and normalize my database.So what I end up doing is creating a new record, then carefully copying over all the fields one at a time (to avoid breaking links to the original entry).

Does anyone else have a better process for updating records?

adamsmith

The current best way is to add a new entry and then use the merge function (available from duplicate items or in the context menu with two items selected) as noted by bwiernik above.

Beyond this, no need to ask for updates -- this is clearly on devs radar and there'll be plenty of posts when it arrives - you won't miss it.

salph

Gonna add my +1 as well.It's probably the single biggest thing preventing me from fully migrating over from Endnote at this point.PDF and website extraction isn't perfect, and manually updating things every time is a significant time sink.

dreikap

188BET靠谱吗+1 from me, just looked for my zotero password to log in here to say that.188BET靠谱吗I understand Zotero's development process is different from other commercial providers, just want to say if this feature comes, it'll be a big improvement

scbarton

+1 for me too.As another convert from Mendeley, this is my most-missed feature.It should not be hard to do since the metadata machinery is already in place.I would be happy to help with this myself if it's possible.

scbarton

There is an open issue on this topic in the repo.The barrier seems to be the possibility of overwriting current metadata, and the fixes suggested are 1) showing the conflict resolution window and 2) implementing an undo mechanism.An argument was also made that there is little advantage to an update mechanism over the current "Add Item by Identifier + Merge Items" approach.

188BET靠谱吗https://github.com/zotero/zotero/issues/1515

I respectfully disagree about the advantages."Add Item by Identifier + Merge Items" requires a copy/paste plus about 8 clicks per item, including a lot of hand-eye movement.A single click, command key or contextual menu item would be much faster, especially if it could be applied to multiple items.

One possible solution is to present a warning dialog if metadata will be replaced, with the option to show the Conflict Resolution window, continue or cancel.This dialog could include a checkbox to suppress all future warnings.I should note that Mendeley updates the metadata silently, and that is my personal preference.

egcavalcanti

+1 for me!This would be really, really useful.I have almost 3000 items with urls/dois/arxiv identifiers but no publication data, that I would like to be able to update metadata for.There is just no way I can do it one by one.Anyone who says there's little advantage in an auto-update simply doesn't have the same problem as most of us on this thread.

I like the solution with the warning dialog, scbarton!Although I would prefer that metadata is never updated silently (or at least not without my consent): Mendeley has screwed up many of my references that way.It should always also give an option to confirm and undo auto-updates, since sometimes this results in error (e.g.the DOI had a typo, and after update every other piece of info about the file that could be used to hunt it down is overwritten).

skiss

+1 This would be amazing.

ventolinmono

+1
Sometimes on first auto update there's no DOI but you can manually add it later and then force recheck to complete the rest of the metadata.

bwiernik

@ventolinmonoYou could check out the DOI Manager plugin, which can retrieve DOIs for you.

ventolinmono

Thanks, but i just tried it on 8 articles and no DOIs were found.Automation may not work always, that's why a manual alternative is needed: pasting DOIs and force metadata recheck.

llity

I need this, some articles miss abstract field.

bjohas

As indicated above, this would be an excellent feature.

Just to add regarding the 'undo'.The some issue exists when merging two items: Once you've merged, the metadata of the other item is gone - so if you made a mistake, that's it.This plugin 188BET靠谱吗https://github.com/edtechhub/zotero-edtechhubimplements (among other things) an 'item history' where the metadata for both items (prior to merge) is written to a note in the merged item (as RIS).Clearly, there are many usability issues with that too, and it's clearly not as simple as an undo - but if you very occasionally lose some data this way, and don't mind the the extra child notes appearing, then it works well.I've found it useful.

In any case, +1 to metadata rechecking!

qqbb

I'm wondering whether the planned update metadata function could "virtually merge" an existing item with a newly retrieved version.Both versions would be kept in the database, but grouped as a single item in the user interface.The virtual item should also appear as a regular item for citation and export functions.(At a more technical level, the virtual item might actually be a merged version.The discarded metadata might be stored as a hidden item, together with some flag for each field to indicate whether it's the new or old version.The hidden item would allow reverting the changes.)

With this, there could be a new Item Review menu next to the existing Locate menu.If the item has been updated, the button could, e.g., show an exclamation mark.The menu would then allow to call the existing merging tool.This Item Review menu could also suggest merging an item with possible duplicate items in the database.

Updating an item could then be independent of the review and merge step, which might happen at some later time.There could be an Updated Items collection for this task, similar to the Duplicate Items collection.

(Some of this might be along the lines of 188BET靠谱吗https://github.com/zotero/zotero/pull/1582.)

bjohas

@qqbb- I would say that in a way, the 188BET靠谱吗https://github.com/edtechhub/zotero-edtechhubdoes this, albeit in an inelegant, lo-fi way: The previous data remains available in a note.The data could be copied from the note onto clipboard, and then 'item from clipboard' would recreate that item.Sure, it's not elegant, but then again, it exists :)