"Incomplete metadata" label on published datasets

[haarli]

We recently set a metadata field to "required", which was "optional" before. Since v 5.14, this leads to the undesired effect that all formerly published datasets without this field show the "Incomplete metadata" label. The label is visible for every user on the published dataset page.

We were wondering if this behaviour is intended. In our opinion, it might be confusing for external users and should only be displayed for draft versions.

Example:
https://edmond.mpg.de/dataset.xhtml?persistentId=doi:10.17617/3.UV4FBN

[pdurbin]

Huh. Probably not! @ErykKul @DieuwertjeBloemen have you seen this?

[ErykKul]

This is not intended. I agree that it should be only visible in draft versions. I will make a pull request for this.

[haarli]

@ErykKul Thank you!

[poikilotherm]

May I suggest to make this behavior configurable and/or visible to a logged in user with permissions on the dataset?

Usually, when a metadata schema changes a field from optional to required, as contributors/curators you might want to see this kind of Information, maybe also be able to make it a fragment to see all things that need attention at once.

[ErykKul]

Thanks, @poikilotherm
As for the fragment, you can already do that in the "My Data" filters:

[ErykKul]

I will make the rule that makes the label visible for all draft datasets and, if not draft, only when the permission is configured.

[lmaylein]

We have a similar problem in 6.1. A dataset that has been published for a long time is no longer listed in the associated dataverse. If you go to the dataset directly, you will find the tag "Incomplete Metadata". We have not changed anything ourselves with regard to the mandatory fields and I cannot find any attribute marked as a mandatory field in the editing mask that is not filled.

Dataset:
https://heidata.uni-heidelberg.de/dataset.xhtml?persistentId=doi%3A10.11588%2Fdata%2F10054&version=1.1
Dataverse:
https://heidata.uni-heidelberg.de/dataverse/iwrgraphics

[lmaylein]

Okay. Apparently we have a problem with these old data sets with the metadata Geographic Bounding Box.

"Geographic Bounding Box has invalid coordinates. East must be greater than West and North must be greater than South. Missing values are NOT allowed."

You only see this when you try to save the metadata again.
But of course it's not so nice that the affected datasets are then silently no longer listed.

[DieuwertjeBloemen]

I believe the addition of validation on the geographic bounding box fields was added quite recently (I believe in 5.13) for some new functionality to do with geospatial information by @pdurbin. We had some errors from that at the time and had to manually correct the records too (@Kris-LIBIS might remember more of the technical details). I think even it remaining in a previous version of the dataset caused issues at the time for us somewhere.

[pdurbin]

"Geographic Bounding Box has invalid coordinates. East must be greater than West and North must be greater than South. Missing values are NOT allowed."

This error and related validation was added in this pull request for 6.1:

• Add validation for entered coordinate values in Geospatial metadata #10142

I hope this helps!

[lmaylein]

Thank you very much. Do you have any tips on how we can find the datasets affected?

[pdurbin]

@lmaylein I don't off the top of my head, sorry. Can you please create a new issue for this? It sounds like we should add something to the 6.1 upgrade instructions.

[mdmADA]

@pdurbin - note that the 'Incomplete metadata' tag is also calculated on the fly for dataset.xhtml and is displayed for the dataset if the metadata is deemed invalid (dataset.xhmtl code, line 146: rendered="{!valid}" ) and for all users. The 'Incomplete metadata' doesn't show on the dataset "fragment" but does show when when the user goes to the problematic dataset.xhmtl page...

For example, some datasets published on ADA's previous Dataverse version had metadata consisting of url's with leading spaces that passed the publication process with the previous Dataset version but are now being picked up in version 6.2 as "not valid" so the "Incomplete metadata" tag is applied.

The tag also shows when the geospatial bounding box values are incorrect as discussed above (and for all users, even logged out) and that previously passed publishing as those are also determined to be not valid.

Also, if there is a mismatch between the policy/license and or Terms of Access (or missing Terms) when a dataset contains restricted files, then the "Incomplete metadata" tag is also applied. [There will also be the message above the dataset title: "Info – Datasets with restricted files are required to have Request Access enabled or Terms of Access to help people access the data. Please edit the dataset to confirm Request Access or provide Terms of Access to be in compliance with the policy."] You have to adjust the licence or add/edit the Terms of Access to fix this.

So the 'Incomplete metadata' is displayed for incomplete metadata, invalid metadata, or a mismatch between licence and Terms of Access (and perhaps other scenarios that I haven't encountered yet).

As the tag is calculated on the fly, and not stored in the database, I haven't been able to determine how to easily identify all the datasets tagged with 'Incomplete metadata'.

[pdurbin]

@mdmADA thanks. I think (and hope) that @ErykKul fixed all this in Dataverse 6.3 with this pull request:

• 10116 incomplete matadata label setting #10172

If anything fell through the cracks, please open a new issue! 😅