Error "Could not save item" (no ris import)

Peter100

I've been trying to get my Endnote library imported with RIS extension. I've altered the RIS library file (see eg discussion: http://forums.zotero.org/discussion/12476?page=1) and it now works for single item export (incl PDF).

However I'm unable to make it work for more than a few references at a time. When I try to export/import the entire library (ca 1500) I initially get the popup error message, "could not save item".

I first ignored this error and decided to let the import continue (the little green process indicator 'importing items' was showing activity). In fact I let the process run the entire night assuming that it might take some time. When I checked it ca 6 hours later it was still running!

After checking the storage folder (e.g. C:\Documents and Settings\plut\Application Data\Mozilla\Firefox\Profiles\gwztlbdl.default\zotero\storage) I realized that nothing had been imported. I tried the import several more times to be sure (only letting it run for a minute or so) but still the same error appeared and still the same green process indicator 'importing items' but no files and no references in Zotero.

I also tried making all the slashes consistent (// \\) but this didn't help.

Three intermediary suggestions to make problems more traceable:

1) include a link to the "faulty" item with the error "could not save item" so it can be found and corrected.

2) on an import error (eg. could not save item) halt the green "importing items" process indicator window so that one is not led to believe that there is still a process that cannot be interrupted.

3) also, give some indication of how much time before the "importing items" process finishes.

Here is one example of a reference that works when imported alone but does not import when included with the entire library import:

TY - JOUR
AU - McCreadie, C.
AU - Tinker, A.
PY - 2005
TI - The acceptability of assistive technology to older people
SP - 91-110
N1 - January
JF - Ageing and Society
VL - 25
N1 - The acceptability of assistive technology to older people
N1 - Ageing & Society (2005), 25: 91-110 Cambridge University Press
Copyright © 2005 Cambridge University Press
doi:10.1017/S0144686X0400248X
Published online by Cambridge University Press 10Jan2005
L1 - file://C:\Documents and Settings\plut\Desktop\temp\endnote_library.Data\PDF\-The acceptabil-1261011968/-The acceptability of assistive technology to older people.pdf
KW - assistive technology; older people; acceptability; felt need; home; housing
N2 - Assistive technology (AT) is defined in this paper as ‘any device or system that allows an individual to perform a task that they would otherwise be unable to do, or increases the ease and safety with which the task can be performed’ (Cowan and Turner-Smith 1999). Its importance in contributing to older people's independence and autonomy is increasingly recognised, but there has been little research into the viability of extensive installations of AT. This paper focuses on the acceptability of AT to older people, and reports one component of a multi-disciplinary research project that examined the feasibility, acceptability, costs and outcomes of introducing AT into their homes. Sixty-seven people aged 70 or more years were interviewed in-depth during 2001 to find out about their use and experience of a wide range of assistive technologies. The findings suggest a complex model of acceptability, in which a ‘felt need’ for assistance combines with ‘product quality’. The paper concludes by considering the tensions that may arise in the delivery of acceptable assistive technology.
AD - Correspondence:
Claudine McCreadie, Institute of Gerontology, Waterloo Bridge Wing, Franklin Wilkins Building, Waterloo Road, London SE1 9NN. e-mail: claudine.mccreadie@kcl.ac.uk
UR - http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=270362#

http://journals.cambridge.org/download.php?file=%2FASO%2FASO25_01%2FS0144686X0400248Xa.pdf&code=7cb41ec7806909cb5d1724b76fbc7feb
ID - 1626
ER -

dstillman

Here is one example of a reference that works when imported alone but does not import when included with the entire library import:

That's unlikely. It's far more likely that there's simply another record that's breaking.

Provide a Report ID.

Peter100

What might that record look like? How could I identify it?

dstillman

Report ID, please.

Peter100

Excuse me, but what is a report ID and how do I find it? I only see the error message pop up and then disappear.

adamsmith

http://www.zotero.org/support/reporting_bugs
(Dan included the link above)

Peter100

since I'm unable to pause/stop the 'importing items' (as mentioned I let it work for more than 6 hours over night and still didn't finish) I'm unable to access the actions menu to check the 'report errors' and generate a report ID. Any suggestions?

perhaps I could email the entire ris file?

adamsmith

how much testing have you done on smaller groups?
I.e. trying to split the file up into 15 groups of 100 entries - see how that goes?

dstillman

I'm unable to access the actions menu to check the 'report errors' and generate a report ID.

Try a Debug ID.

I.e. trying to split the file up into 15 groups of 100 entries - see how that goes?

Or divide by halves.

Peter100

okay... so I started to export in sets of ca 100 as you suggest.

my first set was 74 (no authors and author names that start with 'A'). this time I received this error message:

javascript application: an error occured while trying to import the selected file...

Report ID: 1571180181

hopefully this is something to go on...

dstillman

You have an invalid path. Provide a Debug ID and we might be able to tell you more.

Peter100

The Debug ID is D2145492901.

dstillman

You have an L1 field with multiple paths in it. Search for "1465738528" to find it—it's a Wikipedia page. This is a known issue with the RIS translator.

(Any idea how the field got like that? Was it like that when you exported it or did some other tool you used do that?)

Peter100

I checked that record in Endnote.

I have three URL links and four pdf file attachments in that Endnote record. I put them there once upon a time.

Does this mean Zotero can't handle more than one url link and/or one pdf attachment?

If so, is there is an easy way to comb through Endnote and check for multiple links and/or attachments before export? With 1500 records this is no simple task.

Please let me know if there are any further diagnostics I can run. Perhaps there's a script that could be run on the ris file to cope with the extra urls and/or attachments; perhaps putting them into a separate note or something?

dstillman

Does this mean Zotero can't handle more than one url link and/or one pdf attachment?

It can, but not in that format. I believe Zotero should import multiple files if they're specified on separate L1 lines. The format above is incorrect according to the RIS spec, which requires one-per-line or with semicolon delimiters (though Zotero doesn't currently support the latter).

Peter100

The urls are on separate lines under the URL section in the Endnote record, and the PDFs are separate file attachments under the ATTACHMENTS section of the same record.

Not sure where to go from here. Would you please advise me further (a non-computer whiz)? For instance, is there another find/replace that I could apply in the RIS file before import? For instance any ideas about how to search and replace other LIs with multiple paths?

You mentioned that, " This is a known issue with the RIS translator" but I'm sure I don't know what the "RIS translator" is...

Just to review, my process is (no additional tools):

1) Run Endnote export to create a an output file. This is a TXT file that I rename to RIS (strangely I choose the RIS export format but I can only get Endnote to write to a TXT file which I then must rename to the RIS extension).

2) Open the RIS file in Word Pad or Notepad to perform find/replace: "internal-pdf://" to "file://C:\Documents and Settings\...\PDF\"

3) Import the modified RIS file into Zotero but the fun stops here....

adamsmith

You might be able to achieve this with intelligent search&replace:
essentially you want
L1 - file://...
L1 - file://...
instead of
L1 - file:// ... file://
that should be doable - you might have to run several replace runs -

I'd recommend trying this with a couple of citations.

Peter100

keyword here: "intelligent" ;)

do you mean something like find ".pdf" and replace ".pdf ^l" (manual line break) in a word processor (Word) rather than WordPad or Notepad?

adamsmith

could you post the RIS of said wikipedia here - I'd like to have a look before I start giving more advice.

Peter100

Thank you. Here it is (the lines miss the LI - and UR - tags):

TY - ELEC
PY - 2007
N1 - 16-01-2007
TI - Social Psychology
JO - Wikipedia
PB - Wikipedia.org
N1 - Social Psychology
L1 - file://C:\Documents and Settings\plut\Desktop\temp\master.Data\PDF\Wikipedia - Social Psychology-1465738528/Wikipedia - Social Psychology.pdf
file://C:\Documents and Settings\plut\Desktop\temp\master.Data\PDF\Wikipedia - Social Psychology (psychology)-1985832224/Wikipedia - Social Psychology (psychology).pdf
file://C:\Documents and Settings\plut\Desktop\temp\master.Data\PDF\Wikipedia - Social Psychology (sociology)-1096640032/Wikipedia - Social Psychology (sociology).pdf
file://C:\Documents and Settings\plut\Desktop\temp\master.Data\PDF\-Social psychology-3462227488/-Social psychology.pdf
N1 - blue
N2 - Social psychology is the study of how social conditions affect human beings. Scholars in this field are generally either psychologists or sociologists, though all social psychologists employ both the individual and the group as their units of analysis. Despite their similarity, the disciplines tend to differ in their respective goals, approaches, methods, and terminology. They also favor separate academic journals and societies.

Like biophysics and cognitive science, social psychology is an interdisciplinary area. The greatest period of collaboration between sociologists and psychologists was during the years immediately following World War II (Sewell, 1989). Although there has been increasing isolation and specialization in recent years, some degree of overlap and influence remains between the two disciplines.

UR - http://en.wikipedia.org/wiki/Social_psychology
http://en.wikipedia.org/wiki/Social_psychology_%28sociology%29
http://en.wikipedia.org/wiki/Social_psychology_%28psychology%29
ID - 2057
ER -

It would also be useful to know how add tags like "abstracts" and other information (DOI, etc) in the RIS that Endnote includes in the record but which Zotero puts as attachments. In my opinion this clutters the record. I prefer one record with only PDF attachment(s) and possibly a website image. This helps me locate the records for which I still lack the PDF. Should I start a new discussion with this (related) issue?

Bionatsci

It would also be useful to know how add tags like "abstracts" and other information (DOI, etc) in the RIS that Endnote includes in the record but which Zotero puts as attachments. In my opinion this clutters the record.

"N2" is the abstract tag according to the specification, so if Zotero is importing that as an attached note in your example it's probably a bug in Zotero's import and should probably be fixed.

There doesn't appear to be a DOI tag in the specification (I may have missed it though), do you happen to know what endnote uses for DOI?

Peter100

I believe Endnote uses the tag M3 for DOI (see below).

In my experience Zotero commonly creates at least one superfluous attachment during Endnote RIS import per record. This is a frustration. My guess is that any time Zotero comes across a line without a tag it puts it into a record attachment. Possible exceptions are my example above of lines that begin with file:// or http:// (if in fact these are the source of the invalid path error discussed in this thread) that cause Zotero to malfunction.

I suggest some kind of Zotero script fix (e.g. modified 'ris-fix.vb' ) that assigns a RIS tag to non-tagged lines or tags like N1 or M3 that it doesn't recognize) as a default. I would rather have all the information contained in one record (e.g. DOI) as a miscellaneous text field (e.g. NOTES or ABSTRACT) then to have them spread out into one or more Zotero 'notes' attachment(s).

One possibility is that on initial RIS import into Endnote (from an online database service like Springer) the DOI may get tagged as a note (N1) and put under the NOTES field, in my brief assessment.

The following Endnote RIS export record contains DOI information. The original Endnote record kept the DOI under 'NOTES' but I copied it to their DOI field before export as well for the purposes of this thread. On Zotero import I get two extra attachments: a) doi:10.1080/13691180600751330 and b) June 2006. (Had I changed the L1 field from "internal-pdf://" to "file://" then the PDF should have also imported as an attachment.) In my experience these are common errors. Here is the record:

TY - JOUR
AU - Adkins, Barbara
AU - Smith, Dianne
AU - Barnett, Karen
AU - Grant, Eryn
PY - 2006
TI - Public space as 'context' in assistive information and communication technologies for people with cognitive impairment
SP - 355-372
N1 - June 2006
JF - Information, Communication and Society
VL - 9
IS - 3
N1 - Public space as 'context' in assistive information and communication technologies for people with cognitive impairment
N1 - doi:10.1080/13691180600751330
M3 - doi:10.1080/13691180600751330
L1 - internal-pdf://Public space as '-0311115776/Public space as 'context' in assistive information and---.pdf
KW - cognitive impairment; information and communication technology; ICT; public space; interaction order; habitus; field
N2 - This paper examines emergent issues of `context' raised by the application of information and communication technologies for people with cognitive impairment. The issue of the development and application of cognitive prostheses for this group provides an opportunity to examine assumptions and issues emerging from this area pertaining to understandings of the term `context' in these applications. In this sense the paper takes these assumptions and issues as a point of departure for the development of a `problematic' that can contribute to the study of the experience of cognitive impairment. The paper specifically addresses recent concerns about the lack of knowledge of these experiences in public spaces such as shopping centres, given that this is a critical site for the civic participation of this group. We argue that this participation should be understood in terms of the `meeting of two histories': the history of contemporary requirements governing participation in public space and the habitus of people with cognitive impairment with regard to this participation. The paper proposes that the salience of cognitive impairment in these spaces turns on what it means for individuals to inhabit them as complex `Container Technologies' (Sofia) and underlines the importance of understanding their efforts to attain a sense of normality (Goffman) in these contexts. We propose that this approach can inform research contributing to the development of a `pattern language', informing applications that make cognition a system property in networks that operate between humans, machines and their contexts.
AD - Australasian CRC for Interaction Design, PO Box 2008, Kelvin Grove, QLD 4059, Australia 2: School of Design, Queensland University of Technology, Brisbane, Australia 3: Centre for Social Change Research, Queensland University of Technology, Brisbane, Australia
UR - http://www.ingentaconnect.com/content/routledg/rics/2006/00000009/00000003/art00006
http://dx.doi.org/10.1080/13691180600751330
ID - 1085
ER -

[end of pasted RIS record]

**
So there are two issues that have emerged in this thread still unresolved. First are the 'invalid path' errors caused by PDF and URLs (i.e. lacking L1 or UR tags) and second are the unwanted attachments (as Zotero 'notes') containing small bits of information like DOI, month or issue (spring, winter...) etc. that should be included with the 'new' Zotero record resulting from import.

Peter100

Upon opening Zotero to test the above RIS import in the above post I was able to generate a error report ID (before import): Report ID: 903580716

Perhaps it is related to the above error report (Report ID: 1571180181)

Peter100

Hi Dan Stillman, adamsmith, others,

I'm adding a comment here to bump this up for a final reply:

Any chance to get some follow up on fixes for RIS invalid path errors and extraneous Zotero 'notes' attachments on import from Endnote?