Publishing: When trying to use the overwrite publishing option, error message with failed to publish

[sbarbosadataverse]

Navigate to this dataset and updates the related publication fields to include metadata for "id type" "id number" and "url" field: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi%3A10.7910%2FDVN%2F0C3BTS

Select save and get a successful draft.

Select publish and choose the overwrite option and error message shows up

superuser functionality

Expected that the publish would overwrite the existing version.

Screenshots:

[pdurbin]

This is the "curate new version" feature (superuser only).

@scolapasta @qqmyers and @sekmiller have discussed briefly. @sekmiller unable to reproduce.

Has happened on multiple drafts. Draft still able to be deleted.

Related to inter versions? If one version has a space?

Code assumes previously published version from list. Could it have the wrong version? Line 102. More details in Slack. #8382 is about version differences.

@scolapasta can probably upload the stacktrace.

No actual file metadata changes. Probably from the draft.

Around line 155, curate command is trying to delete links to file metadatas. Watch out for references to file metadatas.

[valentinapasquale]

Hello,
I have just encountered the same error, but I am still working with Dataverse 5.3. Can anyone help?
I have checked the "id" that is mentioned in the error message: I suspected it referred to a file id and that it had been removed for some reasons, but all files are there and there is no file with that id (so it probably refer to another id...). Is it working if I create a new version instead of overwriting?
Thanks for the help!
Valentina

[scolapasta]

@valentinapasquale The removal is a technical aspect of how overwrite works (since what the code for overwrite copies the info from the draft to the last published, then deletes the draft, effectively "removing" the filemetadata objects the draft created). We are still trying to determine the cause of the error, so any additional info you have about that dataset - since it doesn't always happen - would be helpful.

Creating a new version instead should definitely work.

[valentinapasquale]

@scolapasta this is the dataset: https://dataverse.iit.it/dataset.xhtml?persistentId=doi:10.48557/TRGQOD. We want just to correct the spelling of one of the authors' names (Paniello --> Panniello), that was the only change in the draft. That's why I thought overwriting was the best option.
Thanks for the quick and clear answer!

[scolapasta]

@valentinapasquale One other approach you could do is to make the change directly in the DB. It's obviously more risky, but with a relatively simple change like this, shouldn't be too bad. Let me know if you'd like any help to walk you through it.