Import from Mendeley - tags

enury

Hi!

When importing from Mendeley, is it possible to import *only* my Mendeley tags as Zotero tags/marqueurs?

In Mendeley My Tags were the most important thing in the organization of my library. But when I import to Zotero, I get dozens of random tags which I definitely never entered (some numbers which look suspiciously like doi numbers, bits of sentences, others...).

It seems that I have to go through each document and delete the unwanted tag manually one by one. Is there a better solution? Is there an option in zotero to disable the automatic creation of tags?

Thanks for your help!

dstillman

The ones you don't recognize are likely automatic tags from Mendeley — they should show in orange rather than blue if you look in the Tags pane for an item. You can delete all automatic tags at once by selecting "Delete Automatic Tags in This Library…" from the menu in the tag selector in the bottom left.

You can prevent Zotero from adding automatic tags to new items in the General pane of the prefs (though, for what it's worth, automatic tags in Zotero should generally be much higher quality than what it sounds like you had from Mendeley — Mendeley data quality is pretty bad — so if you're at all interested you could try leaving that on for a short time).

enury

Thanks! I did not see a difference of colour orange/blue, but "Delete Automatic Tags in This Library…" worked perfectly!

dstillman

The color difference only appears in the right-hand pane (i.e., the item tags pane, not the tag selector).

bwiernik

@dstillman I find the color difference really hard to see--could the red be supplemented with something like italic text, or automatic and personal tags be listed in separate sections of the pane?

Gurdas_Sandhu

I would vote against using italic text. Separate sections is somewhat better but still not my preference.

I agree that the current color scheme is not easy to notice, though I've got used to it. Maybe just try darker shades of the orange and blue?

WenzheYin

Hi, I have met a similar problem with enury. but mine is a little bit more tricky. some of my manual tags have the same name with the automatic ones after a mendeley importing. the delete automatic tags feature just deletes the manual ones as well if they share a same name with automatic ones. is there a way to get around this?

JS4ntiago

Hi. I am having the same problem. I imported from mendeley, deleted automatic tags, and I am missing many of my own tags. Any possible solution?

jmeluso

I think I may have found a way to keep PDFs and Mendeley tags based on this and other posts on the zotero forums, but it still requires some manual effort (not recommended if you have >1000 entries, for example). Here are the steps I'm using:

1) Export your full library from Mendeley in both RIS and Bibtex formats.

2) Using a text editor (like Notepad++), use find and replace (turn on regular expressions) to do both of the following steps:

Remove all keywords from the RIS file only using the following command in Replace:
Find what, use everything inside these quotes "^KW - .*$/n"
Replace with, leave blank
Save the RIS file.

Remove all keywords from the bib file using the following command in Replace:
Find what, use everything inside these quotes "^keywords = .*$/n"
Replace with, leave blank
Replace all mendeley tags in the bib file using the following command in Replace:
Find what, use everything inside these quotes "mendeley-tags = "
Replace with, use everything inside these quotes "keywords = "
Save the bib file.


3) Back in Zotero, under Edit > Preferences > General, disable "Automatically retrieve metadata for PDFs" and "Automatically tag items with keywords and subject headings"

4) Import both the saved RIS file and the Bibtex file into your library in Zotero.

5) Here's where you'll need the elbow grease. In Zotero, go to "Duplicate items" and merge all of your duplicates. This will take you some time, so following Zotero's recommendation of going back to Mendeley 1.18 may be preferable. If you do choose this method, you'll need to correct any mismatches between citation types which Zotero cannot merge. To do that, Ctrl + click on the source in the merge pair that is *correct* (to only select the incorrect one), and change the Item Type in the Info pane to match the correct entry.

I know this is tedious, and it may not work for everyone, but I spent a lot of time learning this over the past few days. I hope it saves someone else some time!! (And I'm sorry if it doesn't work for you!)

dstillman

@meluso: This is all just to avoid the problem of being able to delete automatic tags after import without losing matching manual tags? Because we can probably just fix that easily. If you transfer via RIS or BibTeX instead of using the official 1.18 approach, you'll still lose various other data (e.g., folders).

jmeluso

@dstillman: Thanks for your response! It's a little more than deleting automatic tags without losing manual tags, though that's part of it. Personally, I don't keep folder structures in Mendeley because I can use the Mendeley-generated bibtex file to refer to whatever I need in LaTeX. But I hear you that the official 1.18 approach keeps other information. I was finding that merging the RIS and BibTeX files gave me the information I needed most: standard citation entires + PDF location + Mendeley tags. The process I described above keeps the PDFs attached to the entry, removes the journal-associated keywords from everything (because personally, they're not as useful as the entries I manually create in Mendeley), and adds the Mendeley tags as Tags in Zotero. If you'd like me to send the RIS and BibTeX files I'm working with as examples of this, let me know! I'm happy to oblige if it helps others!