DOI Registration

172 views
Skip to first unread message

Mike Nason

unread,
Feb 5, 2018, 8:14:02 AM2/5/18
to Dataverse Users Community
Hey all, sorry if this is a rookie question. I'm curious as to when Dataverse registers a DOI. On unpublished datasets, I'm given a DOI. Is that DOI a forecast of what will be registered or does Dataverse register DOIs for unpublished collections?

Context: My institution has DataCite configured, but I want to test with their testing prefix first before I flip the switch to ours. I've only posted an unpublished test set (I don't want to flood a system designed for permanence with test data sets) and it's pretty clear the DOI didn't register. I just want to rule out whether or not the lack of registration is related to our settings and not as intended.

Cheers,
Mike

Stephen Kraffmiller

unread,
Feb 5, 2018, 9:40:51 AM2/5/18
to Dataverse Users Community
Mike,

On Dataset create, Dataverse contacts DataCite to make sure that the randomly generated identifier is not already in use. On Dataset publish, Dataverse again contacts DataCite to actually register the identifier(it also redoes the uniqueness check again, just to make sure). I don't think that any identifiers registered in the DataCite test environment will resolve to the Dataverse url. 

Let us know if you have any other questions

Steve Kraffmiller
Dataverse Team
IQSS

Mike Nason

unread,
Feb 5, 2018, 9:47:06 AM2/5/18
to Dataverse Users Community
Excellent, thanks for the info. That's very helpful. I assume registration happens automatically. If there are errors with registration, I assume there's a system notification somewhere? Is there any way to correct or fix a DOI issue retroactively? 

For example, we have a dataset with a DOI minted using the default settings before we received our credentials. Is there a way to update that record with the proper DOI? I understand fully, by the way, why this wouldn't be built into the UI. I'm not typically one to advocate DOI "corrections". 

Cheers,
Mike

Stephen Kraffmiller

unread,
Feb 5, 2018, 9:58:06 AM2/5/18
to Dataverse Users Community
Yes, the registration happens automatically when you press the "Publish" button. If there are errors with registration then Dataverse will abort the publish process and report the registration error, leaving it to the user to re-try the publish.

Harvard uses EZID for registering DOIs. Through EZID there is an admin console where you can manually create and edit identifiers. Theoretically, it would be possible to convert an identifier from the test environment to production, but as you said - we wouldn't recommend it. It would be safest to recreate the dataset once you've established your production credentials.

Best, 
Steve

Mike Nason

unread,
Feb 5, 2018, 10:07:03 AM2/5/18
to Dataverse Users Community
Great. Good to know. Thanks so much, Steve. Very much appreciated!

Sebastian Karcher

unread,
Feb 5, 2018, 10:50:26 AM2/5/18
to dataverse...@googlegroups.com
DataCite has published a comparable online interface to EZID's called "DOI Fabrica" last October. You should be able to manually make changes there. That said -- you can obviously not change DOIs (they're persistent). If I understand correctly what your situation is -- you have a "pre-existing" DOI and want to have it be reflected in the Dataverse, the only way to do that is by making direct changes to the Database.

We just did this for our existing DOIs and it works pretty smoothly. If that's helpful I can get you the exact database command we ran. The two places where you need to make the change is in the database directly and in the location of the files (which are determined by their PID, i.e. the DOI in this case)

Hope that's helpful,
Sebastian

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.
To post to this group, send email to dataverse-community@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/17d16f3f-da9c-4354-aee2-003655b2a3d1%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.



--
Sebastian Karcher, PhD
www.sebastiankarcher.com

Mike Nason

unread,
Feb 6, 2018, 7:47:21 AM2/6/18
to Dataverse Users Community
Hey Sebastian, 

Good to know about DOI Fabrica. Our issue, in this case, is specifically that we have unpublished content from our early days with Dataverse before we acquired DataCite credentials. Dataverse has, like, a default prefix populated. So we have a handful of unpublished collections with these pre-emptive DOIs. They wouldn't have been registered with any agency, but they're displayed erroneously on the record. 

The one set we'd published in this window we'll just deaccession and do over. 

Database edits might help for the former, though. 

Cheers,
Mike

Sebastian Karcher

unread,
Feb 6, 2018, 9:46:06 AM2/6/18
to dataverse...@googlegroups.com
So do I understand correctly you have an item with a testing DOI and want to give it a real one now that you have a DataCite account? If so see two options:

1. re-ingest the item into Dataverse and delete the original ones (if there are only a few). While you should (almost) never delete items with active DOIs, it's fine to do so for items with testing DOIs. You can do this with a superuser root account in Dataverse using GUI (or, presumably with the same rights using the API). This would seem like the easiest option if you only have a few datasets to correct

2. Mint DOIs for those items using the DataCite API (or Fabrica) and then change them in the database. Here's e.g. what we ran to change 10.5072/FK2/WVES8F (a testing DOI) --> 10.5064/F6QN64NH (a previously generated valid DOI)

The code would obviously need adjustment for your situation, and is without any warranty whatsoever (and you should definitely test any database changes in an easily revertible testing environment first). We tested this both pre- and post publication of the dataset and didn't see any difference

In postgres psql as user postgres:
\c dvndb
UPDATE dataset SET authority = '10.5064' where identifier = '<current dataset 6 char identifier from after the shoulder>';
UPDATE dataset SET identifier = '<new 8 char identifier including F6>' where identifier = '<current dataset 6 char identifier from after the shoulder>';

in file system:
cd /srv/glassfish/dataverse/files/10.5064 (now that it exists - initially had to be created and have its ownership set as below)
mkdir <new 8 char identifier>
mv ../10.5072/FK2/<old 6 char identifier>/* <new 8 char identifier>
chown -R glassfish:glassfish-admin <new 8 char identifier>

curl http://localhost:8080/api/admin/index - to update main page dataset listing







--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.
To post to this group, send email to dataverse-community@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Sherry Lake

unread,
Feb 6, 2018, 3:33:24 PM2/6/18
to Dataverse Users Community
Hi Mike,

Our Dataverse uses EZID to register DOIs, but we have to move to DataCite. So, since it seems that your testing worked...

I want to test DataCite DOI registrations on our test Dataverse server before I change them on our production server. 

Can you send me your DataCite test settings? 

Specifically, this line from the file:/var/local/glassfish4/glassfish/domains/domain1/config/domain.xml
       <jvm-options>-Ddoi.baseurlstring=?????????????????</jvm-options>

And can you send me your Database Settings for: 
  • :DoiProvider
  • :Protocol
  • :Authority
  • :DoiSeparator
Thanks,

 

Sherry Lake | Scholarly Repository Librarian | University of Virginia Library | shL...@virginia.edu | 434.924.6730 | @shLakeUVA | Alderman Library, 160 N. McCormick Road, Charlottesville, VA 22903 | Alderman 563 | LinkedIn Profile | orcid.org/0000-0002-5660-2970  | “Keeper of the Dataverse" 


On Monday, February 5, 2018 at 8:14:02 AM UTC-5, Mike Nason wrote:

Falco KUB

unread,
Jul 23, 2020, 8:41:41 AM7/23/20
to Dataverse Users Community
Hi everybody,

this is not direclty related to the original question, but I didn't want to open a new topic:

Is there any way to check whether DataCite test DOI's have been assigned correctly? I remember that there used to be a website, where you could resolve test DOI's but I cannot find it anymore and the DataCite documentation is a bit confusing.

We have configured test DOI's for our Dataverse Sandbox and they seem to be working nicely there.However, they don't show up in our DataCite Fabrica Test (though I'm not sure, whether they are supposed to?).

Regards,
Falco

James Myers

unread,
Jul 23, 2020, 8:46:48 AM7/23/20
to dataverse...@googlegroups.com

Falco,

The way I check is to login to doi.test.datacite.org (with the same credentials as defined in the Dataverse jvm options) and the DOIs appear there. You’re right that they don’t make the test doi resolvable anymore – that was a nice feature with EZID. There are also API calls you can make to the test servers to see your DOIs, but they should be showing in the Fabrica interface.

-- Jim

--

You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/7444d9f7-f656-40a6-89c7-f859acb05137o%40googlegroups.com.

Sherry Lake

unread,
Jul 23, 2020, 8:50:52 AM7/23/20
to Dataverse Users Community
Hi Falco,

Dataverse DOIs only appear on DataCite (and test datacite) once they are published. Until then the DOIs that Dataverse assigns are just a place holder in Dataverse. The DOI gets registered once published. Is this correct, Jim?

On another note - Jim - Have you been able to log in to the doi.test server at Datacite? I haven't in a couple of months now.

--
Sherry Lake UVa's Dataverse LibraData

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.

James Myers

unread,
Jul 23, 2020, 8:55:39 AM7/23/20
to dataverse...@googlegroups.com

Sherry,

You’re right –the publish step is the first time DataCite sees a DOI in v4.20 and earlier. (In Dataverse 5.0, DataCite DOIs get registered at Dataset creation and I’ve been testing with the pre-releases where you can see the draft DOIs immediately! )

 

I don’t have any trouble logging into doi.test.datacite.org (or the live site) – I’ve been testing over the last few days. Also - I just logged in now using QDR’s GDCC-based test credentials.

 

-- Jim

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.

--

You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/e09931f9-b0b3-4b84-b243-672515998ceao%40googlegroups.com.

Falco KUB

unread,
Jul 23, 2020, 9:41:45 AM7/23/20
to Dataverse Users Community
That's where I was looking, but they don't show up in our Fabrica Test, even though they are 'published' on our Dataverse Sandbox.
No luck with the API either (just returns 404).
Could I be missing something or can I just keep my fingers crossed that it will worked correclty in production?

btw. while creating a manual DOI in the Test Fabrica, I notice that test DOI's can be resolved via https://handle.stage.datacite.org/

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.

James Myers

unread,
Jul 23, 2020, 9:48:01 AM7/23/20
to dataverse...@googlegroups.com

Hmm. Other than checking the doi.baseurlstring, doi.username and doi.password jvm options to make sure those are all the same as when you look manually, I can’t think of anything you can do from the Dataverse end. If those are OK, I would ping DataCite to see what’s going on.

 

w.r.t. the handle URL – I just noticed that as well – seems to be relatively new. If it’s stable (DataCite is going to keep it), we might be able to expose it in Dataverse somehow for test DOIs.

 

-- Jim

 

From: dataverse...@googlegroups.com [mailto:dataverse...@googlegroups.com] On Behalf Of Falco KUB


Sent: Thursday, July 23, 2020 9:42 AM
To: Dataverse Users Community

Subject: Re: [Dataverse-Users] Re: DOI Registration

 

That's where I was looking, but they don't show up in our Fabrica Test, even though they are 'published' on our Dataverse Sandbox.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.

--

You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/5244e1d4-2019-467a-94a1-70c42d292190o%40googlegroups.com.

Philip Durbin

unread,
Jul 23, 2020, 10:28:43 AM7/23/20
to dataverse...@googlegroups.com
Hi Falco,

One thought is that you could try creating a DOI in Fabrica Test using DataCite's API. Assuming that works, you could double check the DOI-related settings in Dataverse. For example, you'll want to make sure :DoiProvider is set to "DataCite" rather than "FAKE". And you'll want the test URL for Datacite: https://mds.test.datacite.org

I hope this helps,

Phil

p.s. Yes, with Dataverse 5, DataCite DOIs will be reserved when the dataset is created. And there are some new APIs for reserving DOIs for older datasets (created before Dataverse 5) that don't have them. We're still hacking on this code a bit but most of it (and the docs) were merged in this pull request if you'd like to take a look: https://github.com/IQSS/dataverse/pull/6901




--
Reply all
Reply to author
Forward
0 new messages