https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=43147
Bug ID: 43147
Summary: Allow a match point to require two independent fields
to both match (e.g. 001 and 003)
Initiative type: ---
Sponsorship ---
status:
Product: Koha
Version: Main
Hardware: All
OS: All
Status: NEW
Severity: enhancement
Priority: P5 - low
Component: Cataloging
Assignee: [email protected]
Reporter: [email protected]
QA Contact: [email protected]
CC: [email protected]
Target Milestone: ---
C4::Matcher's match points cannot express "these two independent fields
must both match" as a single condition. This makes it impossible to
configure a record matching rule that treats a MARC control number (001)
as authoritative only in combination with its control number identifier
(003), which is a common and useful matching strategy for imported
records from external sources that reuse 001 values across different
originating systems.
Currently, adding multiple components to one match point concatenates
their extracted values into a single search phrase (see
_get_match_keys in C4/Matcher.pm), rather than requiring each component
to match independently. This doesn't produce AND semantics even in
principle, because the underlying search indexes for 001
(control-number) and 003 (control-number-identifier) are indexed as
completely separate fields (see
admin/searchengine/elasticsearch/mappings.yaml). A concatenated phrase
search against either index alone can never correctly require both
fields to match, since neither index contains both values together.
Request: add a way to define a match point (or a new kind of
match point) that requires two or more independent search index
values to each be present on the same candidate record, rather than
concatenating them into one phrase search. This would let a matching
rule express "match only if 001 matches AND 003 matches", analogous to
how Koha already supports "required checks" at the matcher level, but
scoped to build a single match point's own pass/fail condition from
multiple independently-matched fields.
Test plan:
1. Create two bibliographic records where record A has 001 = "12345",
003 = "OSt", and record B has 001 = "12345", 003 = "DLC".
2. Create a record matching rule with a single match point on the
control-number index (tag 001) plus a second component for the
control-number-identifier index (tag 003).
3. Stage an incoming record with 001 = "12345", 003 = "DLC" for import.
4. Observe that Koha's matcher cannot be configured to match only
record B (matching both 001 and 003) while correctly excluding
record A (which only shares 001) -- either both records score as
candidates, or neither does, because there's no way to require both
fields to agree for the same candidate as a single condition.
--
You are receiving this mail because:
You are watching all bug changes.
You are the assignee for the bug.
_______________________________________________
Koha-bugs mailing list -- [email protected]
To unsubscribe send an email to [email protected]
website : http://www.koha-community.org/
git : http://git.koha-community.org/
bugs : http://bugs.koha-community.org/
