Question

For my solr implementation I want the query to return the words with and withhout diacriticts regardless if you search with or without diacritics.

To give an example

The search word is “çest” – Solr returns: ‘cest‘, ‘çest‘ and ‘çest ca‘
The search word is “cest” – Solr returns: ‘cest‘, ‘çest‘ and ‘çest ca‘

Currenty the first works. When I search “çest” it returns both cest and çest. However when I search “cest” it returns only ‘cest’

This is how it looks in my schema:

<fieldType name="text_special_search" class="solr.TextField" positionIncrementGap="100">
    <analyzer>
        <filter class="solr.ASCIIFoldingFilterFactory"/>
        <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="\W+" replacement="-"/>
        <tokenizer class="solr.LowerCaseTokenizerFactory"/>
    </analyzer>
</fieldType>

Is there a way to let it work both ways?

Answer 1

If you want either matches, you don’t need the solr.PatternReplaceCharFilterFactory.
This would remove the special characters before they are passed to the ASCII filter.

You can use :-

<fieldType name="text_special_search" class="solr.TextField" positionIncrementGap="100">
    <analyzer>
        <tokenizer class="solr.LowerCaseTokenizerFactory"/>
        <filter class="solr.ASCIIFoldingFilterFactory"/>
    </analyzer>
</fieldType>

You can also use a WhiteSpace Tokenizer to have tokens and use Lower Case as a Filter.

Also, remember the order of execution in an Analyzer is as follows, irespective of the order you have :-

• CharFilters
• Tokenizer
• Filters