I've had this on my mind for a while. I might just get it done this week. Will keep this thread updated

Hello,

Could anyone update me on the issue above. I hope the feature that aurimas talked about will be available soon.

Cheers.

I can see how the volume, issue, pagination metadata can be retrieved if a record already has the DOI or PMID but can this be automated with only the author and title? While PubMed "ePub ahead of print" records are updated after "print", it doesn't add doi strings to existing records that were submitted without them. I have automated re-queries of PubMed to update ePub articles via the PMID but that access requires a license agreement.

I am pleased that some publishers are retroactively adding DOI strings to their backfiles. We have been working to update our records with these retro DOIs. Even with a semi-automated system of performing searches using part of the title, the lead author's name, and the journal name; there is much hand work needed to get to the DOI. The positioning and nature of the DOI display differs across publishers. If the Zotero folks have innovated a way to get the DOI from metadata instead of the other way around I will be mightily impressed.

The plan is to go through google scholar (which already exists as a locate option), but instead of getting the GS citation data (which doesn't have DOIs either, among other problems) to follow the link to the publisher and scrape from there.
Obviously this is a 2nd best to going through the dx.doi resolver to the publisher page, but I'm pretty optimistic that it will generally work nicely.

@ Rintze @ aurimas I knew about the CrossRef utility when I added my message above. In my experience, CrossRef gets it wrong as many times as it gets it right. Are you having better luck?

Hello,

I hope the feature discussed in the original comment will come with Zotero 4.0. I ask Zotero developers to have a look at i,librarian(http://www.bioinformatics.org/librarian/) which contains lots of very interesting features among which the ability to refresh a particular entry. For example, if you have a journal article whose DOI, volume of pages etc is missing or incorrect, with one click you can refresh it and correct or add missing data. That is very cool and important when you have long list of journal articles.

Also, i,librarian integrates scientific literature search from PubMed, PubMed Central, NASA ADS, arXiv, JSTOR® and HighWire Press®. As a chemist, I find this invaluable. I wish if Zotero can have these features....

what is the status of this request? I would really love to have this feature! I am in the process to switch from EndNote to Zotero since I am fed up with the high price and no real updates and useful features.

In the process of switching from endnote to zotero I have discovered that a lot of the endnote entries either have information missing (page number, DOI, url etc.) or have not been imported into zotero. Either way, I was wondering if there was a way to recover the missing information, given that the title is correct, the year, journal etc. I am not sure how one would do that but I am sure there could be a system to find the missing info from a minimal set of existing info about a record.

Cheers,

Lucian

Hi,

Thanks for the quick response! My EndNote library is very old and very big. I have noticed that some information missing such as the page numbers. Anyway I was very excited about how Zotero imported all my references and trying to get used to it and I was thinking that such a feature would be very useful (i don't think I have seen such a feature in a reference manager program).

Well one can only hope that it will happen at some point!

Lucian

I am not fully sure, but I think that here on the forum was post from any developer about possible adding this to some new version in the future.

It is interesting to see that this feature was originally proposed almost 5 years ago! I'd like to add my request for this feature. I have a long list of about 1,000 items sitting in a collection and waiting for the day that I can update these records with missing metadata. Please, please find a way to get this done.

The messages above demonstrate that, while this would be a very nice feature, the several ways to accomplish this are both imperfect and require some sort of negotiating with outside entities. While there are several ways to mostly implement this, at best the current ways to get this to work will at best result in mostly correct metadata. "Mostly correct" isn't good enough for me and I suspect not good enough for the developers. Thus, while this could be done; doing it well will require great and ongoing effort.

The NLM has essentially so relaxed their license rules that a license really isn't needed. But to retroactively gather missing volume, issue, and pagination requires an identifier of some sort.

CrossRef has improved its search capabilities and opened up many of the restrictions on access.

Hovever, without an identifier (DOI, PMID, etc.) this becomes much more complicated to implement and any implementation will risk a mismatch and incorrect metadata being added.

One example of a best case scenario is the problem with updating PubMed records that began as ePub ahead of print. Some publishers volume, issue, pagination metadata includes the updated "print" date but many publishers do not update the year from the ePub year. This leads to a mismatch between the volume number and the publication year. Specifically, all issues of volume 27 were released in published issues in 2017, but as some of the articles included in volume 27 were initially released as ePubs in 2015 and 2016 the year will not match the volume. (CrossRef seems to avoid this problem.)

Note that some publishers still do not update the publication date on their own website's metadata but at the same time include the article within an issue of a volume published a year or two after the article was initially released.

As @aurimas said above this title search (without an identifier number) can be implemented with Google Scholar queries and a follow through to a publishers website. But some percentage of the GS best results will be incorrect matches. And some of the correct matches will lead to publishers that provide incorrect information.

All of this presumes that the interval between when the item was added to your Zotero library and the search for updated metadata has been sufficiently long enough for the item to have been added to a published issue.

I have (not Zotero) database systems set up so that I get updates from publisher ftp sites, the NLM for PubMed, and CrossRef. But I also have a logic test that examines publication volume and publication year. (This is made more complex because some journal volumes run July through June or September through August.) Even with this automation, it remains necessary to have a human take a few moments to look at and think about any metadata update. Just when I think that the system can be set on full-automatic; a publisher implements an unusual idea that breaks my automated system.

One more thing about PubMed metadata updates. Elsevier example: All Elsevier journal volumes are not now divided into issues. When articles are added to the PubMed database as ePub records the system nonetheless includes an ePub flag for a non existent issue. This must be accounted for when the metadata is updated but there is still no issue. There will be a need to replace the issue-ePub-flag with nothing.

That may not have been clear: The problem is really not what to do in theory. There are some decisions to be taken that affect the details, but on a technical level this is eminently doable. The problem is that there are 300 feature requests and 3 developers.

Dear developers!

I appreciate your effort to maintain and update Zotero. And this feature should be super useful for everyone.
Hope it would be paid more attention and set to the nearest updated version of Zotero, or, at least, there is a plug-in to do this.

Wish you all good health!
Thank you very much!