Morning All,
I have been using the following guide
to add some big data files that failed validation to a dataset. I accidentally transposed one with the wrong storage identifier and md5 that belong to a different file.
How do I safely edit/delete this so I can get everyting added to dataset.
Sherry left us some great notes but this is not covered.
Tim
Well, I don't think you want to simply delete the file from the Dataverse UI. From a quick test, it'll delete the file from S3. :grimacing:
I'm playing around with this test: mvn test -Dtest=S3AccessIT#testDirectUpload
And seeing what shows up in LocalStack at http://localhost:4566/mybucket
Jim Myers is the expert here but he isn't on Zulip.
@Tim Stevens would you mind posting to https://groups.google.com/g/dataverse-community ?
Morning Phil. Sherry's notes indicated that deleting it from within the UI would require a re-upload of the file so will try to avoid that since the person uploading this 3D data is so far behind I dont want them to have to backtrack with a re-upload. Would updating the persistent identifier and md5 in the DB table and re-indexing also be a workable solution? Seem low friction unless there is another element that needs updating that I am not aware of...
If you can find the right table, yes, I think that would work.
Is some sort of API needed here? Seems like an honest mistake that should be correct-able somehow. :thinking:
The API may not be a bad idea. If the deletion within the Admin UI gave the option to remove the DB record and if desired keep the physical file in S3 that would also work.
Well, please feel free to make a feature request if you like. :smile:
OK, will do. I am a command line guy but the folks they are interviewing to replace Sherry are not so it seems like its worth requesting.
I'm a command line guy too :smile:
Last updated: Oct 30 2025 at 06:21 UTC