Stream: community

Topic: Update a file without changing major version


view this post on Zulip odelaigue (Oct 24 2024 at 12:32):

Is there any possibility to update a dataset file (a README file in my case), and to avoid to have a change of a major version?

I think that in that case it doesn’t make sense because a README file is  more or less a metadata, and the data didn’t change…

I would have liked to have changed only the minor version number (or without changing version).

view this post on Zulip Vera Clemens (Oct 24 2024 at 12:56):

We have also thought about this.

E.g. for cases where the file is changed only because a typo within the file content has been fixed, a minor version increase might be enough.

Based on my brief look at the code and some testing, I don't think it's possible though.

view this post on Zulip Philip Durbin 🚀 (Oct 24 2024 at 13:04):

I'm afraid it isn't. Not for changes to the content of files. If the checksum (e.g. md5) has changed, we force a new version.

view this post on Zulip Philip Durbin 🚀 (Oct 24 2024 at 13:05):

There's somewhat related functionality if the typo is in a metadata field like "description". A superuser can fix it and republish without bumping the version using an "Update Current Version" button. Please see https://guides.dataverse.org/en/6.4/admin/dataverses-datasets.html#make-metadata-updates-without-changing-dataset-version

view this post on Zulip Philip Durbin 🚀 (Oct 24 2024 at 13:06):

But again, for files, we force a major version change.

view this post on Zulip Philip Durbin 🚀 (Oct 24 2024 at 13:06):

Should we treat README files differently? Have different rules for them? :thinking:

view this post on Zulip odelaigue (Oct 24 2024 at 13:15):

Thank you for your answer. Yes, I have seen that there is no problem about changing metadata without updating the version.
In my opinion, the README file (or any explanation file, that is files that don't have the data) should be viewed as similar to metadata.
I think that is confusing to the users to upgrade the major version number when there is no change in the data (e.g. for a typo as said Vera Clemens previously).

view this post on Zulip Philip Durbin 🚀 (Oct 24 2024 at 13:21):

Have you played with adding tags like code, data, or documentation to files? Maybe these could be signals to Dataverse to treat files differently.

view this post on Zulip odelaigue (Oct 24 2024 at 13:28):

It seems to have the same result for all tag options. It forces for a major version change.

view this post on Zulip Philip Durbin 🚀 (Oct 24 2024 at 13:29):

Yes, sorry, I mean in the future, as some sort of new feature.

view this post on Zulip odelaigue (Oct 24 2024 at 13:33):

I think that it could be a very good solution!
If the owner of the dataset uses a "documentation" tag for a file, he can then decide between major or minor version change.

view this post on Zulip Philip Durbin 🚀 (Oct 24 2024 at 14:13):

Cool, I suspect it'll be a fair amount of work to implement this, though. You don't happen to have any developers who would want to poke around in this part of the code, do you? :sweat_smile:


Last updated: Nov 01 2025 at 14:11 UTC