lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Anders Nielsen" <and...@visator.dk>
Subject RE: prefix query with multiple words
Date Tue, 04 Dec 2001 13:09:21 GMT
I've made a "hack"-solution for this. It basically makes a BooleanQuery with
alot of OR-branches. Each OR option corresponds to a complete phrase, and
like in the code for PrefixQuery I take the last term in the phrase I want
to search for and make a TermEnumeration and find all the terms that has the
search-term as the prefix. For each of those I make a complete PhraseQuery.

A solution where it was possible to add an array of terms instead of a
single term, to a PhraseQuery would most likely perform alot better.


-------------------------------------

public
class PhrasePrefixQuery
{
    public static Query getQuery(IndexReader reader, Term[] terms)
    {
        Term prefixTerm = terms[terms.length-1];
        TermEnum enum = null;

        BooleanQuery result = new BooleanQuery();

        try {
            enum = reader.terms(prefixTerm);

            do {
                Term term = enum.term();
                if (term != null &&
term.text().startsWith(prefixTerm.text()) && term.field() ==
prefixTerm.field()) {
                    PhraseQuery pq = new PhraseQuery();
                    for (int i=0;i<terms.length;i++) {
                        if (i == terms.length-1)
                            pq.add(term);
                        else
                            pq.add(terms[i]);
                    }

                    result.add(pq, false, false);
                }
                else
                    break;
                }
            while (enum.next());
        }
        catch (IOException e) {
            e.printStackTrace();
        }
        finally {
            if (enum != null)
                try {
                    enum.close();
                }
                catch (IOException e) {
                    e.printStackTrace();
                }
        }

        return result;
    }
}

-----Original Message-----
From: Tom Barrett [mailto:barrett_tom@yahoo.com]
Sent: 4. december 2001 00:42
To: lucene-user@jakarta.apache.org
Subject: prefix query with multiple words


Hey all-

Wondering if it's possible to a prefix query, but with multiple words;
basically trying to get

+artist:"eric clap*"

to return documents with artists "eric clap", "eric clapton", "eric
claptonean", etc.

You can get close by parsing into multiple words first and prefixing the
last word (i.e. "Eric Clap" -> +artist:eric +artist:clap*), but this also
gives you results that have the phrase in the wrong order (i.e. returns
results with artist "clap eric")

Is there any way to do this right?

Thanks,

Tom

 _________________________________________________________ Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
--
To unsubscribe, e-mail:
<mailto:lucene-user-unsubscribe@jakarta.apache.org>
For additional commands, e-mail:
<mailto:lucene-user-help@jakarta.apache.org>



--
To unsubscribe, e-mail:   <mailto:lucene-user-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:lucene-user-help@jakarta.apache.org>


Mime
View raw message