Bug 33401 - Diacritics are ignored when auto-creating authorities
Summary: Diacritics are ignored when auto-creating authorities
Status: NEW
Alias: None
Product: Koha
Classification: Unclassified
Component: MARC Authority data support (show other bugs)
Version: 22.05
Hardware: All All
: P5 - low normal
Assignee: Bugs List
QA Contact: Testopia
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2023-04-04 09:15 UTC by Klas Blomberg
Modified: 2025-03-31 13:12 UTC (History)
1 user (show)

See Also:
GIT URL:
Change sponsored?: ---
Patch complexity: ---
Documentation contact:
Documentation submission:
Text to go in the release notes:
Version(s) released in:
Circulation function:


Attachments
Different subject heading in same authority record (163.52 KB, image/png)
2023-04-04 09:15 UTC, Klas Blomberg
Details
screenshot of linked auths (110.26 KB, image/png)
2025-03-31 13:12 UTC, Thomas Klausner
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Klas Blomberg 2023-04-04 09:15:29 UTC
Created attachment 149122 [details]
Different subject heading in same authority record

Koha doesn't take diakrits in consideration when authority records are created.

In swedish we have three extra letters in the alphabet - ÅÄÖ. Koha treats those letters as AAO, and creates one unified authority record for the subject headings "Ankor" (ducks) and "Änkor" (widows)

So, if you search for widows as subject you will get matches for Donald Duck.
Comment 1 Katrin Fischer 2023-04-04 09:17:28 UTC
I assume this might be because it appears the same to the search engine/SQL. Maybe an additional check is needed before creating the authority.
Comment 2 Thomas Klausner 2025-03-31 13:03:38 UTC
This seems to be fixed via Bug 38729.

After applying the patch and setting the new syspref LinkerConsiderDiacritics to 'do', when linking via "Link authorities automatically", "Ankor" and "Änkor" are correctly linked.

Not sure though what exactly you mean by this:

> So, if you search for widows as subject you will get matches for Donald Duck.

Where do "search for widows"?

If I don't autolink, but click on the manual search / linking popup thingy, I get two results (Ankor and Änkor). And can manually then pick the correct one.

But again, I assume that 38729 fixes this bug.
Comment 3 Thomas Klausner 2025-03-31 13:12:20 UTC
Created attachment 180030 [details]
screenshot of linked auths

Ah, taking the title of this bug as a hint, I tried:

* setting  AutoCreateAuthorities to 'generate'
* editing a biblio
* adding a new 650$a 'Töst'
* adding another new 650$a 'Täst'
* save
* click on the 'autolink button'
* the 650s point to two different auths
* (see screenshot)