Citekey duplicates on iOS?

[iandol]

I have quite a few recent papers where my citekeys are duplicates, like I have two [huang2024] citekeys. Desktop bookends seems to append a letter so I think the culprit for this is iOS bookends. Can you check if the citekey that is created on iOS is supposed to handle duplicates (i.e. if huang2024 exists, make the new one huang2024a)?

[Serge]

Thank you for reporting this! This has now been fixed and will be included in the next update.

[iandol]

Great, thanks for the awesome support as always!

[iandol]

Note: for anyone who has duplicate keys, I found the easiest solution was to export as a BibTeX and use JabRef which has a duplicate keys tool: Quality > Check Integrity to get the list of duplicates and then change them manually

[Jon]

Assuming you are syncing the iPad to Bookends Mac, you can select all references (Command-A) or just the problematic ones and use Refs -> Generate Citekey. This will update the citekeys for the selected references and ensure there are no duplicates.

Jon
Sonny Software

[iandol]

Perhaps I'm mistaken but i noticed this could sometimes rename my oldest citekey, so for example say I have:

doe2024 doe2024a doe2024

Selection and regenerating got me

doe2024a doe2024b doe2024c

i.e. the first two changed. I did check this ages ago, then just developed a superstition not to use generate citekey on older citation, perhaps I am mistaken / was fixed?

[iandol]

I just tested on desktop, lets say I have this situation (sorted by added date, this seems most relevant), two duplicates added later by bookends iOS:

Screenshot_2025-06-12-11-47-45.png (4.77 KiB) Viewed 53 times

Generate citekey yields:

Screenshot_2025-06-12-11-48-06.png (5.07 KiB) Viewed 53 times

It seems sorted by author alphabetically (first name), and every citekey is renamed, I would rather try to preserve the existing ones and add letters based on how recently they were added to the database:

Screenshot_2025-06-12-11-52-49.png (5.01 KiB) Viewed 53 times

This means if I have cited zhang2022 or zhang2022a in a paper they will still be valid...

[Jon]

The citekeys are regenerated de novo, and of course if there are duplicates the one that is encountered first gets the primary citekey and the other is disambiguated. If one or more is used to create temp citations in a paper being written, then yes, you'd have to ensure that the temp citations for those references still pointed to the right reference. You could exclude such "active" references from the citekey refresh. For example, if they are in a group, you could create a smart group that would find all references NOT in that group, select all (Command-A), and generate citekeys for those. But maybe that's more trouble than it's worth.

Jon
Sonny Software