My Mendeley import included ~6,000 automatically-added labels. Is there a way to batch delete them?
Perhaps exporting from Paperpile and processing in some other software like JabRef?
We don’t have batch functions for labels at the moment. I had to think long to find a workaround but that might work:
- Temporarily move all papers in the Trash. That can take some time, wait until it’s done.
- All your labels should now be “unused”, i.e. no number right to it. Click “Clean up”. That will delete all unused labels.
- Restore the papers from Trash.
I’m interested a bit more in the reason why you have so many labels? Where does Mendeley get them from? Are these keywords added by the authors and converted to tags?
Perhaps @andreas knows more and we can fix the root cause. I know for Zotero import we only convert user generated tags to Paperpile labels and ignore author keywords. Not sure about Mendeley.
This is a very clever solution and worked perfectly! It does take time: ~1.5 hours to move to trash 3,000 references, one day to delete ~6,000 labels, 1 minute to restore paper from trash (folder structure preserved!).
I am not sure about the source of these labels. I did go through a Zotero import, they might have been introduced there.
Thanks for the brilliant tip!
Typically trashing is and restoring is relatively fast. But once the things you are trashing have many labels things get very slow. We need to fix this at some point.
The Zotero import might explain it. If your papers happen to be from PubMed Zotero adds a tag for each “author keyword” that comes with the paper.
This is still an issue when importing from Mendeley. Perhaps a way to make set a threshold on how many times the tag is currently used when cleaning up the tags. For example: delete all tags with less than 3 uses.
Your solution could work, but I have many tags that I actually use joften. The ones automatically imported by Mendeley have very few other uses and could be removed as stated above.
I am really upset at Medeley for this…
" Where does Mendeley get them from?"
— looks like they are manually added by the Mendeley user.
Hi, related to this post: is it possible to turn off the default labeling? I would really appreciate this functionality since I use customized labels only. The solution in this post seems to be a temporary solution, since we have to go through this remove-all-labels business every once in a while…
I have the same problem as the OP (coming from a zotero import, which originally came from mendeley). The proposed solution sounds like it takes a day, and involves temporarily moving papers to Trash, which is terrifying. This was in 2015. Is there no better solution?
What about adding something similar to the cleanup button, but allowing a threshold? Even removing labels with 1 citation would be useful.
It would also be useful to trash all labels and start again. They are completely useless to me now as there are so many, if I accidentally expand then I have to scroll multiple pages to get to my shared folders.
Is API access an option? I would be willing to write a script to do this.
+1 on API access. Given the pace of development, opening up a portion of the API could be an efficient way to let the users themselves deal with the recurrent annoyances around the metadata.