Prefix query is not getting analysed.

[modassar81]

Hi Tim,

With current Lucene-addon when I try to search fl:"Sto* books" with following analysis chain I see the prefix query not changing to sto* as expected.

Analysis chain. <tokenizer class="solr.WhitespaceTokenizerFactory"/> <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt"/> <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="1" catenateNumbers="1" catenateAll="1" splitOnCaseChange="1" preserveOriginal="1"/> <filter class="solr.ICUFoldingFilterFactory"/>`

During debugging I found the normalise method of Analyzer class is not calling the classes of above analysis chain.

Can you please help?

Best, Modassar

[modassar81]

Hi Tim,

Can you please provide your suggestion? It is getting a blocker and I am not able to migrate to newer Solr version. Kindly look into it.

Best, Modassar

[modassar81]

Hi Tim,

I tried to understand the logic of lucene-5205 and noticed that the normalize() method is not using the analysis chain. When I looked into older lucene-5205 I found that it uses following code for prefix queries.

protected Query newPrefixQuery(String fieldName, String termText) throws ParseException { Analyzer mtAnalyzer = getMultiTermAnalyzer(fieldName); if (mtAnalyzer == null) { return handleNullAnalyzerPrefix(fieldName, termText); } BytesRef analyzed = analyzeMultitermTermParseEx(mtAnalyzer, fieldName, termText); return wrapMultiTermRewrite(new PrefixQuery(new Term(fieldName, analyzed))); }

So I tried to use the older lucene-5205 code with Solr-8.6.3 and found that it is very slow for a query which used to work very efficiently with Solr-6.5.1. The other queries are fine. I understand this can be due to some differences in Solr-6.5.1 and Solr-8.6.3. Please correct me if I am wrong.

Kindly help.

Best, Modassar