Page MenuHomePhabricator

Subcategories are only partially sorted (multiple "runs")
Closed, ResolvedPublic

Description

On the abovementioned category page, the subcategories are sorted only partially. I.e. the list of initial letters looks like "A,B,C,D,E,F,I,A,B,E,M,N,P,S,U,V,Z,Ř,L,N,P,R,S,T,Č,Š". That is, three separate runs in the list.

Wild (!) guesses of the problem cause:

  • on some pages, [[Category:Evropa]] is used, on other [[Kategorie:Evropa]].
  • diacritics (Ř, Č, Š)

(Note: Our Special:Version reports 1.3.5+, which is not in the above list.)


Version: unspecified
Severity: normal
URL: http://cs.wikipedia.org/wiki/Kategorie:Evropa

Details

Reference
bz616

Event Timeline

bzimport raised the priority of this task from to Medium.Nov 21 2014, 6:58 PM
bzimport set Reference to bz616.
bzimport added a subscriber: Unknown Object (MLST).

This is a known problem with MySQL. See bug #164.

  • This bug has been marked as a duplicate of 164 ***

jeluf wrote:

The "multiple runs" are caused by problems with the sortkeys in older revisions
of the
MediaWiki software. The DB has now been cleaned up using

update categorylinks set cl_sortkey= substring(cl_sortkey,11) where cl_sortkey

like 'Kategorie:%';

New entries should have the correct sortkey.

The sorting of non-latin characters is indeed a mysql localization issue.

My bugreport did not reference the MySQL sorting issue (which is known to me,
and I take it as not critical, although somewhat unpleasant). So that this bug
is indeed not a duplicate of #164.
But because JeLuF solved the problem, all is fine now. :-)
Thanks to JeLuF.

byrial wrote:

We also have this problem in several categories in the Danish Wikipedia. See
http://da.wikipedia.org/wiki/Kategori:Danskere for an example. According to
JeLuf's comment, it can fixed by a simple database command. Who should be asked
to kindly do that?