I would assume it's just sorting by UTF8 decimal order like most programming
languages do. See:
https://www.w3schools.com/charsets/ref_utf_basic_latin.asp
https://www.w3schools.com/charsets/ref_utf_latin1_supplement.asp
I believe you'll have to either store a new field with a custom sort (if you
want to handle it inside the Lucene index) in it or you'll need to write a
custom sort function.
On 2/1/22, 5:58 AM, "Sebastian Riemer" <s.rie...@littera.eu> wrote:
Hello,
I have documents with family names in german language context.
The field type is defined like this: <fieldType name="string"
class="solr.StrField" sortMissingLast="true" />
When sorting by this family name, I get results ordered like that:
Ascending order:
1. Bart
2. Otz
3. Ozzbourne
4. Zacharias
5. bariton
6. biene
7. burtsch
8. ozza
9. Ádele
10. Òle
11. Ônna
12. Ötz
13. ägnie
14. órthega
So, the ordering is:
1. A-Z (Uppercase)
2. a-z (Lowercase)
3. Umlauts and special accents
That ordering seems suprising to me, I'd prefer it like this:
1. aäàâ-z (Lowercase)
2. AÄÀÂ-Z (Uppercase)
Or in other words, lowercase before uppercase, and Umlauts and Special
accents after its "natural character".
Here is the full query fort he above example:
start=0
&rows=50
&fq=tenant_id:1
&fq=u_markedAsDeleted_b:false
&fq={!tag%3Du_cg_customergroup_0}(((u_customerGroupMemberships_customergroup_cp_ts_ns:(24))))
&fq=u_id_cp_s:[*+TO+*]&q=*:*
&facet=true
&facet.missing=true
&facet.sort=count
&facet.mincount=1
&sort=u_familyName_cp_s+asc,u_userName_cp_s+desc
&qf=u_userName_cp_s^20+u_displayName_cp_s^20++text^2+text_en+text_de+text_it
&pf=u_userName_cp_s^100+u_displayName_cp_s^20++text^10
&mm=100%25
Could you please give me some direction which parts of the documentation I
can study to learn about sorting and how to achieve a custom sorting?
Thank you and best regards,
Sebastian
