[leomoon]

4 years, 2 months ago #13927

When searching using diacritical characters, ćčšđž, the search should also hit csdz as secondary search results. This works well on quicksearch results, but not on full results.

On my site at https://test.antikvarijatknjiga.hr you can try with Andric (also returns Andrić, as it should), Dorde (Đorđe), Sibenik (Šibenik), Zivotic (Životić), Covjek (Čovjek).

[wpsolr]

4 years, 2 months ago #13930

This is related to search engine configuration, more precisely analysers.

Special characters of a language can be treated by specific language filters.

For instance:

– https://lucene.apache.org/solr/guide/6_6/filter-descriptions.html#FilterDescriptions-ASCIIFoldingFilter or https://lucene.apache.org/solr/guide/6_6/charfilterfactories.html#CharFilterFactories-solr.MappingCharFilterFactory
To replace characters with their Lating form: “á” => “a”

– Hungarian stemmers like https://www.solr-start.com/javadoc/solr-lucene/org/apache/lucene/analysis/hu/HungarianLightStemmer.html

….

[leomoon]

4 years, 1 month ago #14463

Thank you. I added

<filter class=”solr.ASCIIFoldingFilterFactory” preserveOriginal=”false” />

to
<analyzer type=”index”>

and

<analyzer type=”query”>

sections and now all characters are rendered down properly.

[wpsolr]

4 years, 1 month ago #14465

Great. Thanks for the feedback.