Mailing-List: contact java-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-user@lucene.apache.org
Received-SPF: pass (asf.osuosl.org: local policy)
Message-ID: <42F1F16A.2040204@programmfabrik.de>
Date: Thu, 04 Aug 2005 12:43:54 +0200
From: Martin Rode <martin.rode@programmfabrik.de>
User-Agent: Mozilla Thunderbird 1.0.2 (Windows/20050317)
MIME-Version: 1.0
To: java-user@lucene.apache.org
Subject: Re: Search in multi fields with cross field AND ?
References: <013F0E81D349EC42A14CD3242BC61AFFF9A96B@sch-mail-06.schibsted.no>
In-Reply-To: <013F0E81D349EC42A14CD3242BC61AFFF9A96B@sch-mail-06.schibsted.no>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Erlend,

try this code:

QueryParser q = new QueryParser("text", analyzer);
q.setOperator(QueryParser.DEFAULT_OPERATOR_AND);
           
Query query = q.parse(search);
Hits hits = isearcher.search(query);
 

Best,
Martin


erlend.bjorge@aftenposten.no wrote:

>Hi there!
>
>I'm fairly new to Lucene and trying evaluate it. I want to retrieve only documenets (companies) that has 
>all the words searched for in different fields, crossed in different fields with AND search.
>
>I have a structure in the "documents" (companies) like this :
>
>companyName string
>keywords1 string        (Extra boosting if match here)
>keywords2 string
>
>---
>
>So far so good when I create the index with boosting on keywords1 etc.
>
>If I search for:
>
>"ford garage"
>
>I want companies that rapair ford cars, and only those.
>
>So if you have documents/companies like:
>
>"company1", "ford volvo", "garage motor ...."
>"company2", "volvo nissan", "garage ......."
>"company3", null, "ford garage motor ....."
>
>---
>
>I will have match with company1 and company3. Company2 don't do ford cars. Company3 haven't match in keywords1, but 
>they still doing garage stuff with Ford cars, but should be after company1 in the result list, since match in keywords1 
>should give you extra boosting = higher score = why I don't have only one field for both keywords.
>
>So I have tried different approaches doing searching, something like this:
>
>1 ---------------------------------------------------------------------------
>
>        String[] fieldNames = new String[]{"companyName", "keywords1", "keywords2"};
>        String searchText = "ford garage" / "ford AND garage"
>
>        BooleanQuery query = new BooleanQuery();
>
>        for(int i = 0; i < fieldNames.length; ++i) {
>            QueryParser qp = new QueryParser(fieldNames[i], new StandardAnalyzer());
>
>            qp.setOperator(QueryParser.DEFAULT_OPERATOR_AND);
>            query.add(qp.parse(searchText), false, false);
>        }
>
>        Hits h = is.search(query);
>
>2 ---------------------------------------------------------------------------
>
>        Query query = MultiFieldQueryParser.parse(searchText, fieldNames, new StandardAnalyzer());
>
>        Hits h = is.search(query);
>
>------------------------------------------------------------------------------
>
>In both cases, I only get company3 in the result because "ford garage" is within the same field for this company I guess, 
>and I don't get Company1 too. I guess because the two words are split in two different fields = search in muliti fields, 
>cross field AND..... if you understand ?
>
>So is this possible with Lucene, what have I missed ? I have the Lucene book, can't find anything there, tried to find
>anyting in the archive, didn't find anythign there either, but maybe I didn't look good enough ? What should I look for ?
>
>So any clues of how I can do this search ??
>
>
>Thanks in advanced!
>
>
>Cheers, 
>Erlend 
>
>  
>


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org