Report ID: 526320099 (Migrating from Mendeley)

sarahfaulwetter

Hi all,

I am trying to migrate from Mendeley to Zotero and keep running into errors during the import.
My library is large (ca 20.000 entries, of which ca. 6500 have attached files).
I do have Mendeley 1.19, but my sqlite file is somehow not encrypted, probably I've copied a backup over into the Mendeley directory after upgrading to 1.19.

The first few hundreds to thousand records are imported correctly, then an error pops up:
Warning: NetUtil.newChannel(uri) deprecated, please provide argument 'aWhatToLoad'

Warning: NetUtil.newChannel(uri) deprecated, please provide argument 'aWhatToLoad'

[JavaScript Error: "Translator Crossref.js with same ID is already loaded -- deleting CrossRef.js" {file: "file:///home/user/Zotero/translators/CrossRef.js" line: 0}]

[JavaScript Error: "Translator Crossref REST.js with same ID is already loaded -- deleting Crossref-REST.js" {file: "file:///home/user/Zotero/translators/Crossref-REST.js" line: 0}]

[JavaScript Error: "IndexedDB Maintenance finished with error: NS_ERROR_NOT_AVAILABLE: ActorsParent.cpp:18877"]

[JavaScript Error: "TypeError: x.fileURL.endsWith is not a function" {file: "chrome://zotero/content/import/mendeley/mendeleyImport.js" line: 114}]

version => 5.0.55, platform => Linux x86_64, oscpu => Linux x86_64, locale => en-US, appName => Zotero, appVersion => 5.0.55, extensions => Zotero LibreOffice Integration (5.0.14.SA.5.0.55, extension)


This is the third time I tried, I had emptied the folder /home/user/.zotero/zotero/ and deleted the folder /home/user/.mozilla/firefox/xxxxxxxxx.default/zotero before each new attempt.


Anything I can do from my side to get my records and the file paths over? Manually via bib /ris export and then running a script or direct sqlite queries to update file paths?

Thanks a lot.

dstillman

Sorry about that — it looks like there's an unexpected value in your Mendeley database for one of your file attachments. Can you try the latest Zotero beta and see if that helps? If you get another error, provide another Report ID.

No need to delete anything before retrying.

sarahfaulwetter

Hi all,

Follow up to yesterday's problem / Error Report which is described here: https://forums.zotero.org/discussion/73477/report-id-526320099-migrating-from-mendeley
(I opened a new discussion to provide a new Report ID)

Tried to import Mendeley library again, this time using the beta version, but the same error occurs:

Warning: NetUtil.newChannel(uri) deprecated, please provide argument 'aWhatToLoad'

Warning: NetUtil.newChannel(uri) deprecated, please provide argument 'aWhatToLoad'

[JavaScript Error: "TypeError: x.fileURL.endsWith is not a function" {file: "chrome://zotero/content/import/mendeley/mendeleyImport.js" line: 114}]

version => 5.0.56-beta.12+4b64cc312, platform => Linux x86_64, oscpu => Linux x86_64, locale => en-US, appName => Zotero, appVersion => 5.0.56-beta.12+4b64cc312, extensions => Zotero LibreOffice Integration (5.0.14.SA.5.0.56-beta.12+4b64cc312, extension)


I guess I'll go through my Mendeley records and see if there are any corrupt file links. Any other suggestions?

Thanks.

dstillman

Actually, if you're comfortable doing some lower-level debugging, we can do this a bit more efficiently. If you open your Mendeley database in either the SQLite command line client or a GUI tool and run SELECT * FROM files WHERE TYPEOF(localUrl) != 'text';, do you see any results?

If you're still getting an error in the beta, that might mean that there's a URL that's actually a number. I can add another test for that, though I'm curious what the value is.

dstillman

@sarahfaulwetter: The latest beta should also avoid this, regardless of what's in those fields (though I'm still curious).

sarahfaulwetter

Thanks, @dstillman . No results from the sql query, but in the meantime I had manually cleaned up my Mendeley files by deleting broken links and non-pdf files. So there might have been a result from the query but I've probably removed that file by hand.

I am rather sure, however, that none of the files / links I removed was or ended with a number. These broken links and non-PDFs were not many and I eyeballed them all for strange patterns. Only thing I can imagine is an (invisible) linebreak in the filename, e.g. at the end.

The latest beta I tried did not solve the problem, but manually removing stuff from mendeley did, at least I have imported now > 10,0000 records instead of ~ 1,000.

Thanks!