Import from bib file fails

bhaddow

Hi

I'm trying to import a bib file with approximately 1000 entries. However it fails after about 10% of the entries, reporting that there was an error with the file. Is there any way to get more information about which entry failed and why?

cheers - Barry

adamsmith

provide an error report ID:
http://www.zotero.org/support/reporting_bugs
Best way to find out why import fails is to split the file in chunks

bhaddow

Hi Adam

The error report id is 166366548. I tried to imported my entire citeulike file (id: bhaddow).

I guess what you're suggesting is that I do something like a binary search to find the offending record(s). I was hoping that zotero would be able to provide a helpful debug message somewhere to avoid such a tedious search,

cheers - Barry

aurimas

see first section here http://www.zotero.org/support/debug_output

instead of submitting to Zotero server, you can view output.

bhaddow

Hi Aurimas

Thanks, I found the error log. However it doesn't appear to report the problem entry. Do you want me to post the log?

cheers - Barry

aurimas

The log is probably quite large. It should have been submitted for the report ID that you provided, so no need to post the actual log (though, if you wish, you may post it on a website like gist.github.com or pastebin.com). The Zotero core developers will have to take a look at the report you submitted, which may take a few hours.

bhaddow

Hi

So after having a better look at the logs, I discovered that zotero was objecting to invalid url fields in my bibtex file. After several rounds of delete url / rerun / examines logs, I eventually go zotero to process the complete file without errors. This took about 4-5 hours, and resulted in 188 out of about 950 entries being loaded into my zotero account. Not very successful.

cheers - Barry

noksagt

How were the URL fields invalid? Invalid URLs that are still valid BibTeX should largely work. Have you run your file through, e.g. bibcheck/bibclean to confirm there aren't other issues?

bhaddow

Hi

I did not check them all, but I think the URLs failed because the documents they pointed to had moved. When I removed the offending URLs, the file appeared to process to completion - although most of the entries were lost.

I tried bibcheck, but it just reported lots of spurious errors then segfaulted. Using the file in a latex document through bibtex works fine,

cheers - Barry

noksagt

Feel free to post the failing file somewhere. But....

By 'URL' field, do you then mean either 'file' or 'pdf' (or 'sentelink')? Yes, Zotero will try to import and attach a file associated with these links & missing files may prove problematic.

bibcheck is a reasonably robust program. If it is segfaulting, I can imagine there might be a reasonably significant problem with your file. You might also try using 'bibutils' to convert it. If that also fails, I'm not sure if it will be reasonable to expect Zotero to ever handle it.

latex+bibtex doesn't usually check all references and fields the way that importing into other bibliographic software or conversion of bibcheck. Unless you have created a tex and bst file to thoroughly test everything, that's fairly anecdotal & doesn't surprise me terribly.

bhaddow

Hi Noksagt

The offending fields are 'pdf' fields. It's reasonable for Zotero to object to dangling pdf links if it wants to import them, but a more specific error message would be nice. I've no idea why it still failed when I removed these links, but since it tied up my browser for a couple of hours, I'm reluctant to try again.

The errors that bibcheck reports are mainly non-standard fields and formatting errors in the 'comments' field.

Since importing a bib file is proving problematic, is there any other way to get data from citeulike?

cheers - Barry

noksagt

Assuming you use the same username, the RIS file works for sure (I just imported it). There is also a CiteULike site translator that should work (though may require a bit more labor on your part). I'm about to test BibTeX.

noksagt

After stripping pdf links from the BibTex file, it imported 958 records & did so quickly. Did you make other changes to the file?

noksagt

CUL options: don't prefix keys, prefer personal key; otherwise use numeric key, Exclude Amazon links, Don't escape URLs, smart wrap whole field.

But the RIS export is probably better for this.