[Koha] Authority linking problems with 650 records (TOPIC_TERM, local thesaurus)

Doug Kingston dpk at randomnotes.org
Mon Nov 8 07:59:43 NZDT 2010


As Elaine pointed out in her earlier letter, we basically solved the
our initial authority linking problem by fixing the 008 field, but
this only fixed the "PERSO_NAME" authorities.  Our TOPIC_TERM
authority records were still not getting linked.  These Names and Term
authorities are locally maintained since the "regular" thesaurus's one
might use like Library of Congress do not have the level of detail and
specificity that this specialist resarch library requires
(http://www.efdss.org/front/library-archive/about-the-vaughan-williams-memorial-library/37).

We have this working now, but there were two issues.   First and
easiest to fix was that we were placing our TOPIC_TERMs in a 653
record, rather than a 650.  C4::Heading::MARC21 has a table of
bib_heading_fields, and 653 is not included.  We researched 650 a bit
and realized that we probably should just use that, but we wanted to
be correct and not claim these terms were sourced from someplace like
LC.  So we created a 650, with a second indicator of '7' (other), and
subfields 'a' set to the term, and '2' set to our library identifier
'UkLoVW'.  When we tried to use link_bibs_to_authorities.pl, it failed
to link these terms.  Some debugging and generous printfs indicated
that the problem was in the query against the Zebra index.
Specifically, the problem was with the line:
          $limiters .= " AND Subject-heading-thesaurus=$self->{'thesaurus'}";
Commenting out this line (and thus the constraint that the entry in
Zebra have a thesaurus of 'UkLoVW') solved the problem.
Query before change:
 authorities:SimpleSearch(Match-heading,ext="Aberdeenshire" AND
at='TOPIC_TERM' AND Heading-use-subject-added-entry=a AND
Subject-heading-thesaurus=UkLoVW)
Query after change:
 authorities:SimpleSearch(Match-heading,ext="Aberdeenshire" AND
at='TOPIC_TERM' AND Heading-use-subject-added-entry=a)

I don't know if this was the right thing to do and I suspect if I put
in a patch to delete that line, it would be rejected.

Here is an example (from marcdump) of one of our terms.  This MARC was
generated by a python program I wrote to convert from a bespoke term
database to MARC so there could definitely be issues with this which I
would happily fix.  Comments are welcome.

LDR 00234nz## 2200109n######
001     AT-000001
003     UkLoVW
005     20101106220637.0
008     101106|ge|dz||aaan##########||||c###|d
035    _aAT-000001
040    _aUkLoVW
       _beng
       _cUkLoVW
150    _aAberdeen

The real problem is why did Zebra not have our entries label with the
correct thesaurus entry?  Thiis code in C4::Heading/_query_limiters()
looks a bit suspect anyways with lots of "FIXME" comments by someone.
So where is the real problem - does Zebra need fixing?  Our authority
MARC records?  Koha C4::Heading?  What should we do to resolve this
permanently?

Regards,
    -Doug-


More information about the Koha mailing list