EBSCO Academic Search Premier--Everything's a Journal

ln108
I've just grabbed twenty or so references from Academic Search Premier. Working OK, except that it identifies everything as a journal article: book chapters, even entire books.

Curious. Anyone have any ideas?

Thanks!
  • adamsmith
    that's the data coming from EBSCO - if you're curious you can use EBSCO's export function and look at the downloaded .RIS file and you'll find
    TY - JOUR
    in all of them. Not much Zotero can do. I'd suggest sending EBSCO an e-mail.
  • Keith Ralston
    Sent them an email.

    I also have EndNote which I believe reads the RIS. It is capable of looking at the data and sorting between journals and book sections.
  • adamsmith
    huh - that's interesting about endnote. Wonder how that works.
  • ln108
    edited November 3, 2010
    Thanks, adamsmith. I sent EBSCO an email. We'll see what comes of it.

    Meanwhile, if ENDNOTE can sort this, why can't Zotero?
  • noksagt
    that's the data coming from EBSCO - if you're curious you can use EBSCO's export function and look at the downloaded .RIS file and you'll find
    TY - JOUR
    Not always. I searched for 'Smith' & enabled only books/monographs to match. The first listing is: Virginia Beach, Virginia, 2007.Full Text Available World Almanac & Book of Facts, 2009, p375-375, 1p and is typed with:TY - GENCould it be something as simple as EndNote & Zotero using different fallbacks for the generic type?

    ln108: you'd have to look at the RIS for your particular export to see if this was the case for you too.
  • adamsmith
    yes - sorry. When I initially wrote this I confused this with Wilson Web, which has the same problem and indeed uses wrong RIS code.
  • Keith Ralston
    Oddly enough. If I choose Direct Export to EndNote, ProCite, CITAVI, or Reference Manager while sitting in Zotero; the item imports in Zotero properly. In contrast, Zotero's URL icon imports the record as a Journal Article.

    hmmmmm
  • dstillman
    I haven't studied it, but Zotero's EBSCO translator has some fixes for bad RIS data before it passes it to the RIS translator. It's possible EBSCO has cleaned up their data and those fixes are no longer necessary.
  • ajlyon
    The translator includes this line:
    if (!text.match(/TY\s\s-/)) text = text+"\nTY - JOUR\n";
    Perhaps EBSCO is leaving more or less space after the TY tag and the type tag is being added regardless of whether it is already specified? I'll explore this.
  • ln108
    I have the following response from EBSCO tech support on this issue:

    After checking with our product managers, I have found that we have a current service issue open in regards to the inability to export citations to Zoltero. I have attached your case to this service issue and will let you know as soon as it has been resolved.

    I do apologize for the inconvenience and will let you know of any new progress to this issue. Please let me know if you have any questions in the meantime.
  • ajlyon
    Actually, EBSCO already contributed code to make it possible to export citations to Zotero. This bug might actually be on Zotero's side -- I'm having trouble accessing EBSCO to test it myself, but you can try out a potential fix.

    Please go to http://github.com/ajlyon/zotero-bits/raw/master/EBSCOhost.js and save the file to the translators directory of your Zotero data directory (http://www.zotero.org/support/zotero_data).

    Reload the EBSCO item page and see if it will save correctly.
  • noksagt
    One immediate error I caught:"TypeError: text.match(/DO\s+\-\s*(.*)/) is null"Presumably because the export containsKW - OBERNDORF, Meyera E.. The regex should probably match from the beginning of a line.
  • noksagt
    EBSCO does type (at least many) newspapers and magazines as journal articles. The books/mongraphs that I've found are all typed as GEN (which should import into Zotero as 'document'). It had sounded as if ln108's sources might be different. The github version of the translator does still type them as journals, though.
  • ajlyon
    And that's not even something I changed. That's fixed in Github. Can you check again?
  • noksagt
    Presumably because it still has the line item.itemType = "journalArticle";?
  • ajlyon
    The culprit is line 125:
    item.itemType = "journalArticle";
    Can you try again now with that line commented out (available on Github now)?
  • ajlyon
    Right... I was looking for something like the WilsonWeb problem, and I jumped on the first possible reason. Without working access to EBSCO, it's a little hard to test.
  • ajlyon
    The DOI error reported by noksagt is fixed in the Github version. If someone can install it and test the book section issue, we can see about checking in a fix to the repository.
  • noksagt
    Yeah, this works for items typed as GEN. The DOI error is also fixed, though I don't know why we look for those three terms anywhere in a line (rather than at the beginning of a line).
  • ajlyon
    I've changed the term lookup to look at the beginning of the line; the new version is on Github. Testing by people with working EBSCO access is appreciated.
  • ln108
    ajlyon,

    Using the new js file, I've successfully grabbed a book section as a book section, and a book as a book. Will report any further glitches. Many thanks!
  • ajlyon
    This is now in the repository (https://www.zotero.org/trac/changeset/7301).

    It will go to clients in the next Zotero release or when Dan pushes it. Hopefully this can be pushed, since the current version handles all non-journal-article items incorrectly.
Sign In or Register to comment.