lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Erick Erickson <erickerick...@gmail.com>
Subject Re: IndexSearcher
Date Sun, 08 Mar 2009 14:09:18 GMT
In Luke, you can select from among most of the analyzers provided by
Lucene, there's a drop-down on the search tab. And I'm sure you can
add your own custom ones, but I confess I've never tried.

Best
Erick

On Sun, Mar 8, 2009 at 4:14 AM, liat oren <oren.liat@gmail.com> wrote:

> Ok, thanks.
>
> I will have to edit the code of Luke in order to add another analyzer,
> right?
>
> If I need to query a long text - for example to search for articles that
> are
> textually close in their content, I need to parse to  the query the text of
> the article. Then I get the error that it is too long. Is there any way to
> solve it?
>
> Thanks a lot,
> Liat
>
> 2009/3/5 Erick Erickson <erickerickson@gmail.com>
>
> > I think your root problem is that you're indexing UN_TOKENIZED, which
> > means that the tokens you're adding to your index are NOT run through
> > the analyzer.
> >
> > So your terms are exactly "111", "222 333" and "111 222 333", none of
> which
> > match "222". I expect you wanted your tokens to be "111", "222", and
> "333",
> > each appearing twice in your index.
> >
> > Try indexing them tokenized. Although note that I don't remember what
> > StandardAnalyzer does with numbers. WhitespaceAnalyzer does the
> > more intuitive thing, but beware that it doesn't fold case. But it might
> be
> > an easier place for you to start until you get more comfortable with what
> > various analyzers do.
> >
> > Also, I *strongly* advise that you get a copy of Luke. It is a wonderful
> > tool
> > that allows you to examine your index, analyze queries, test queries,
> etc.
> >
> > But be aware that the site that maintains Luke was having problems
> > yesterday,
> > look over the user list messages from yesterday if you have problems.
> >
> > Best
> > Erick
> >
> > On Thu, Mar 5, 2009 at 8:40 AM, liat oren <oren.liat@gmail.com> wrote:
> >
> > > Hi,
> > >
> > > I would like to do a search that will return documents that contain a
> > given
> > > word.
> > > For example, I created the following index:
> > >
> > > IndexWriter writer = new IndexWriter("C:/TryIndex", new
> > > StandardAnalyzer());
> > > Document doc = new Document();
> > >  doc.add(new Field(WordIndex.FIELD_WORLDS, "111 222 333",
> > Field.Store.YES,
> > > Field.Index.UN_TOKENIZED));
> > > writer.addDocument(doc);
> > > doc = new Document();
> > > doc.add(new Field(WordIndex.FIELD_WORLDS, "111", Field.Store.YES,
> > > Field.Index.UN_TOKENIZED));
> > > writer.addDocument(doc);
> > >  doc = new Document();
> > >  doc.add(new Field(WordIndex.FIELD_WORLDS, "222 333", Field.Store.YES,
> > > Field.Index.UN_TOKENIZED));
> > >  writer.addDocument(doc);
> > > writer.optimize();
> > >  writer.close();
> > >
> > > now I want to get all the documents that contain the word "222".
> > >
> > > I tried to run  the following code but it doesn;t return any doc
> > >
> > >  IndexSearcher searcher = new IndexSearcher(indexPath);
> > >
> > > //  //  TermQuery mapQuery = new TermQuery(new Term(FIELD_WORLDS,
> > > worldNum)); - this one also didn't word
> > > Analyzer analyzer = new StandardAnalyzer();
> > > QueryParser parser = new QueryParser(FIELD_WORLDS, analyzer);
> > >  Query query = parser.parse(worldNum);
> > >  Hits mapHits = searcher.search(query);
> > >
> > >
> > > Thanks a lot,
> > > Liat
> > >
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message