Allow to export metadata of unpublished datasets

[rmo-cdsp]

Hello !

I was wondering why we cannot export metadatas of unpublished datasets, while we can with published datasets ? For a matter of quality control, this functionnality would be useful.

[pdurbin]

@rmo-cdsp good question. Which format do you want? Dublin Core? DDI? JSON? If you want JSON, a workaround would be to download metadata from the "native" API while passing in an API token of a user who has access to the unpublished dataset.

[rmo-cdsp]

@pdurbin It would be on "regular" user side, so using the native api may not be that good. And they are also more into DC and DDI than JSON. To me, the mechanism already available for published datasets would be the best.

[jggautier]

I emailed @rmo-cdsp to get more detail about the use-case behind this issue, "matters of quality control," and he (very quickly) replied that

they want to be able to check the exported data with their own, in case they did not match because of human error, or by Dataverse itself

Researchers who are concerned about how their metadata is being mapped to different standards (which can affect how the data is described in different systems, affecting discoverability) want to see the mapping before they publish the dataset. If the mapping isn't what they expect, they can adjust how they're interpreting Dataverse's metadata fields and reenter metadata or raise concerns that the mapping might be wrong.

One open question is what are the benefits of doing this mapping check on the draft page, before the dataset is published? Why not publish, and then check and adjust?

[pdurbin]

concerned about how their metadata is being mapped to different standards

Makes sense. Reminds me a lot of this issue: Connect metadata field names and blocks to (de facto) standard ontologies #2357

they want to be able to check the exported data with their own

@rmo-cdsp how are your researchers doing this check? What tools are they using?

Why not publish, and then check and adjust?

My impression over the years is that researchers and curators like to get the dataset just right the first time they publish their 1.0 version. Attitudes on this differ, I'm sure. Consider how Nick Lauland from TDL who said, "I'd like to be able to fix this typo as if it never occurred, but I'm guessing that opens a whole ball of wax as far as preserving all changes, etc." (emphasis mine) over at https://groups.google.com/d/msg/dataverse-community/HYMHkrQ0dUI/1qreu8kABAAJ

[BPeuch]

I second this request. It is something we wished we had during our own testing phase at the State Archives of Belgium.

[cmbz]

To focus on the most important features and bugs, we are closing issues created before 2020 (version 5.0) that are not new feature requests with the label 'Type: Feature'.

If you created this issue and you feel the team should revisit this decision, please reopen the issue and leave a comment.