On 7 Jan 2009, at 5:31 AM, Chris Borst wrote:
> Thanks for your help Adam. Yes, that worked, by and large. My
> apologies
> for not finding it in going through the Help docs. Absolutely, the
> issue
> is multivalued fields.
>
> The "by and large" reflects two continuing, but entirely workable,
> problems:
>
> (1) As a multivalued field, only the first and last items in the list
> are bounded by { or }, so {ARISTOTLE shows up separately from
> ARISTOTLE.
> (2) Many references also contain their DOI numbers, which contain
> periods and so end up parsed as (multiple) groups. Though I haven't
> yet
> changed the default setting to '. ' to see what effect that has.
>
I'm not sure WHICH pref you want to change, I expect
BDSKDefaultGroupFieldSeparatorKey? But I wouldn't do that anyway. It
certainly won't work. As I said, we currently hard-code ";" for the CR
field in the ISI parser anyway. Also, this won't affect the SPLITTING
of group values, which is determined by
BDSKGroupFieldSeparatorCharactersKey, and which is interpreted as a
set of characters rather than a single string.
> There's also the obvious challenge that many references to the same
> item
> end up grouped separately, either due to ISI's sloppy data entry
> practices, authors sloppy citation practices, or quite proper
> citations
> of different editions or pages of a work. But those are easily
> enough met.
>
This sloppyness can never be fixed automatically by BD, you should do
that by hand.
> Just for future reference, a Cited References field looks like this:
>
> Cited-References = {{{*}DOH, 2001, NHS PERF RAT AC TRUS.
> {*}DOH, 2002, NHS PERF RAT INTR.
> AKKERMANS HA, 2005, J OPER RES SOC, V56, P931, DOI
> 10.1057/palgrave.jors.2601923.
> BARKER RM, 2004, BRIT MED J, V329, P107.
> BOHRNSTEDT GW, 1994, STAT SOCIAL DATA ANA.
> BOLLEN KA, 1989, WILEY SERIES PROBABI.
> CARVEL J, 2003, GUARDIAN 0716.
> CASTANEDAMENDEZ K, 1998, J HEALTHCARE QUALITY, V20, P1.
> FITZPATRICK M, 2002, NUR MAN, V33, P35.
> GRIFFITH JR, 2002, J HEALTHC MANAG, V47, P1.
> GULLAND A, 2002, BRIT MED J, V325, P236.
> HOYLE RH, 1995, STRUCTURAL EQUATION.
> JORESKOG KG, 1989, LISREL, P7.
> KAPLAN RS, 1992, HARVARD BUS REV, V70, P71.
> KAPLAN RS, 1996, BALANCED SCORECARD T.
> KMIETOWICZ Z, 2003, BRIT MED J, V327, P184.
> KOSSOVSKY MP, 2000, AM J MED, V109, P386.
> LEYLAND AH, 1995, IMA J MATH APPL MED, V12, P175.
> MANNION R, 2005, J HLTH SERV RES POLI, V10, P18.
> PATEL B, 2005, FORWARD BACKWARD STE.
> SHIFRIN T, 2003, GUARDIAN 0702.
> SNELLING I, 2003, J HLTH ORG MANAG, V17, P210.
> SOLANO J, 2003, INFORMATION SYSTEMS, V20, P66.
> SORBOM D, 1989, PSYCHOMETRIKA, V54, P371.}}
>
There's a lot of commas in there, so grouping will plit on those as
well. You can try using ";" for the
BDSKGroupFieldSeparatorCharactersKey hidden pref.
> and is always followed by:
>
> Number-of-Cited-References = {{24}}
>
> It contains the complete bibliography of every article in the ISI
> database (the sample is a shorter one!).
>
> I haven't yet tried to do such a search from BibDesk's internal ISI
> search feature.
>
BTW, if you have downloaded the ISI bibliography, you can also just
drop it on the main table in BD, and let BD do the parsing. This will
also clean the CR field by inserting ";" between the lines. Thinking
about that, we should probably use the default grouping separator.
Although...
> Thanks again,
> Chris
Christiaan
------------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It is the best place to buy or sell services for
just about anything Open Source.
http://p.sf.net/sfu/Xq1LFB
_______________________________________________
Bibdesk-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bibdesk-users
