Hi,

I'm in a similar position. Want to use zotero but my primary citation database is in bibtex and I want to import it rather than regenerate it in zotero.

Because you picked a good thread name I'd like to propose we use this thread to document the issues we hit with bibtex importing so that they are more visible to developers. We should restrict discussion to the mapping of bibtex data to fields in the zotero database as much as possible. Would you agree?

Based on some other specific threads:
http://forums.zotero.org/discussion/2107/bibtex-book-chapter-title/
It might be worthwhile some of us writing out a mapping of bibtex fields to zotero fields so that writers of import and export filters can have a resource to work from?

Here is another example of an issue in waiting:

BibTex@inproceedings:
Required fields: author, title, booktitle, year
Optional fields: editor, pages, organization, publisher, address, month, note, key

Zotero:Conference Paper
Doesn't provide separate fields for conference organiser and proceedings publisher (bad) and separately stores the "conference name" which will typically be highly redundant with the proceedings title (confusing); there is also no provision for editor information (import ->export may lose important information).

Small collection of related threads:
http://forums.zotero.org/discussion/1950/
http://forums.zotero.org/discussion/2134/
http://forums.zotero.org/discussion/2125/
http://forums.zotero.org/discussion/1952/
http://forums.zotero.org/discussion/1904/
http://forums.zotero.org/discussion/1789/

Here is today's installment: a conference paper.

@inproceedings{pkey,
author = {First von Last and von Last, First and von Last, Jr, First},
title = {paper-title},
booktitle = {proceedings-title},
year = {2008},
editor = {von Last, Jr, First},
pages = {200--205},
organization = {Organiser},
publisher = {Publisher},
address = {Address},
month = {Month},
note = {note, useful for storing the authors affiliations},
key = {key},
abstract = {not in bibtex ``standard'' but quite common},
keywords = {also in common use},
doi = {will become important},
leftoffield = {some other data we are not going to worry about}
}


Import problems
--------------------
1. First von Last name format is not correctly converted. The von part is grouped with the first name.
2. Names with a Junior part are not correctly converted. The first name is lost
3. latex style double quotes are not converted to as ascii or UTF replacement
(as above)
4. Conference organiser (organization) is lost (field is not present in zotero Conference paper entry)
5. a single key-phrase is broken into words (comma separated phrases are not)


Discussion
-------------
Problems 1-3,5 are shared with other publication types and represent general bibtex import problems. In this case doi information is preserved.

Significant to this publication type is 4: the fact that zotero does not provide storage of the conference organiser. It is not always the same as the organiser. Springer is an example of a publisher that publishes proceedings without organising conferences.

Export problems: (sorry should put them elsewhere but I have them right now)
---------------------
1. Proceedings title is exported to "journal" but should be exported as "booktitle"
2. Month is not exported unless it is recognised as a real month name. A better behaviour would be to export as a string.

The second point is again a "general" issue

So to take stock here is a list of bibtex import and export issues that should (?) have a bug/ticket/whatever raised for them. I have left out a couple of smaller issues that, in retrospect, may not be bugs (such as no doi for reports):

1. Zotero author data fields insufficient to handle names containing "von" and "Jr" parts
2. Author information is lost when importing from bibtex where names contain von and Jr parts
a. "First von Last" name format is not correctly converted. The von part is grouped with the first name.
b. "von Last, Jr, First" name format is not correctly converted. The first name is lost (and Jr used in its place)
3. latex double quotation marks are not correctly imported from bibtex. eg. ``text''. The start and end quote marks comprise 2 ascii characters and need to be converted to an appropriate unicode character within zotero. (may be an issue in export also but not tested)
4. When importing from bibtex, if the keyword field contains a single key-phrase, it is broken into single words in zotero. Import should probably always assume comma delimited keywords to prevent this.
5. Institution is lost (field is empty in zotero) when importing from a bibtex @techreport entry
6. When exporting a Report entry to bibtex, an @misc entry is generated and should be a @techreport. Possibly as a consequence, the report number is lost,
7. zotero does not provide storage of the conference organiser in the conference paper entry type. Consequently, conference organiser (organization) is lost when importing from bibtex @inproceedings
8. When exporting a conference paper to bibtex, the proceedings title is exported to "journal" but should be exported as "booktitle"
9. When exporting a journal article to bibtex, "issue" should be exported "number" but is not resulting in loss of information
10. When exporting page ranges to bibtex and en-dash "--" should be used but a single dash "-" is used.

I'll stop at that for a while unless anyone asks me to generate test data etc for any other specific bibtex entry types.

I've patched 3,5,6,8,9,10.

Thanks heaps. That's truly amazing turnaround :-)

Take the point on 4. Keywords is non-standard in bibtex anyway

Re 1,2 I understand the complexity of this. I've been using last,first in all my bibtex because it is the least ambiguous and results in fewer formating errors, especially in cases where bibliography styles dictate short form names. I think in the longer term 4 part names will probably be required in zotero but ... like I said I understand its a relatively complex implementation issue. (The link I posted on names previously in this thread is useful)

7 is not going to bite me any time soon. Its more for completeness.

Stoked!

When will 1.05 come out BTW?

yes... had to remove duplicates. Might be just a problem with me not using the latest repository version.

At lest I know where to go now to fix things :-)

In the diff view here:
https://www.zotero.org/trac/attachment/ticket/1020/scrapers.diff

the new / corrected entries look like they are out of order in the table. I guess that's where my confusion arose. And yes I was looking at the zotero->bibtex table... so double confusion.

I didn't have or post a test case where the zotero entry contained the unicode quotation characters but again my expectation would be that they would map to the latex equivalent.

Having patched my tables manually (in the most obvious way) it seems to behave perfectly.

Only readability as far as I know. I personally had not encountered the verbose form so you are starting to stretch my knowledge as a latex user. I had initially assumed it was a XML hybrid thing but it looks like it is a newer latex thing. Looks sensible in fact! And is probably easier for other applications to parse/import/export than the obscure latex forms. It's probably good to leave as is.

Can the importer handle both the short and long forms though?

Follow-up: I notice that my Logos-Zotero export-import issue relates to items # 6 and 8 in andre's list.

Your post doesn't really relate to the thread. The issue with the export from Logos. BibTeXing doesn't list journal as an optional field for inbook or for misc.

There are a few thread on the Logos support site about this. This one has a link to a way of fixing their BibTeX export and also a suggestion on a work around:
https://community.logos.com/forums/p/103224/713491.aspx