[Exist-open] Query String Optimization within Lucene's Standard Analyzer

[Jochen G. <joc...@un...>]

Hi,

a question concerning ft:query (describing the Query with XML),
the StandardAnalyzer and Lucenes Query String Optimization:

Having a XML with punctuation like this:

<text>
    <string>string. string</string>
    <string>string-string</string>
    <string>string_string</string>
    <string>string, string</string>
    <string>string/string</string>
</text>

Querying it by:

//string[ft:query(., 'string')]
or
//string[ft:query(., 'string/')]
or
//string[ft:query(., 'string-string')]

... always leads to 5 results because Lucene's Standard Analyzer
and obviously the Query Parser too ignore / filter out punctuation.

If one describes Queries in XML (a feature which has great possibilities
in my eyes), obviously the Query Parser does not do Optimizations
any more.

//string[ft:query(., <query><term>string.</term></query>)]
or
//string[ft:query(., <query><wildcard>string-*</wildcard></query>)]

... always returns an empty sequence.

One exception is the "Phrase-Search":
//string[ft:query(., <query><phrase>string-string</phrase></query>)]
which returns 5 results as above.

So my question is: Is it planned from your side to add this Query
Optimization to the XML based Fulltext Query on a low level or
am I responsible for this optimizations myself within my application?



Best,
Jochen


Tested with:
eXist-setup-1.4.0-rev10440.jar

<collection xmlns="http://exist-db.org/collection-config/1.0">
    <index>
        <fulltext default="none" attributes="no"/>
        <lucene>
            <analyzer 
class="org.apache.lucene.analysis.standard.StandardAnalyzer"/>
            <text qname="string"/>
        </lucene>
    </index>
</collection>