lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From iouli.golova...@group.novartis.com
Subject Re: MultiSearcher to Indexing.
Date Fri, 13 Aug 2004 12:03:50 GMT

Well, actually we use a nice piece of hardware with a lot of memory and 2
cpu under linux.

As front-end we use coldfusion application.  Seems to be ok, but we have
not  tested on huge load yet. Let You know if smth. gettig wrong.

Regards,
J.




                                                                                         
                                             
                      "Vladimir Yuryev"                                                  
                                             
                      <vyuryev@rambler.        To:       "Lucene Users List" <lucene-user@jakarta.apache.org>
                         
                      ru>                      cc:                                    
                                                
                                               Subject:  Re: MultiSearcher to Indexing.  
                                             
                      13.08.2004 13:06                                                   
                                             
                      Please respond to        Category:       |-------------------------|
                                            
                      "Lucene Users                            | ( ) Action needed       |
                                            
                      List"                                    | ( ) Decision needed     |
                                            
                                                               | ( ) General Information |
                                            
                                                               |-------------------------|
                                            
                                                                                         
                                             
                                                                                         
                                             




Hi Joel,

Parallel method requests a lot of memories, but MultiSearcher requires
slightly less memory.
Tomcat at the large loading gives out a system mistake.
If you have other experience of work that please tell me.

Regards,
Vladimir.

On Fri, 13 Aug 2004 12:22:34 +0200
  iouli.golovatyi@group.novartis.com wrote:
>
>Hi Vladimir,
>
>Can You please explain me what's the benefit of this approach and why
>_pickles_?
>
>I f I understand correctly the ?-n was how to make query run
>paralelly on
>multi-index. Is ParalelMultiSearcher not for this?
>
>Regards,
>Joel
>
>
>
>
>
>                      "Vladimir Yuryev"
>
>
>                      <vyuryev@rambler.        To:       "Lucene
>Users List" <lucene-user@jakarta.apache.org>
>
>                      ru>                      cc:
>
>
>                                               Subject:  Re:
>MultiSearcher to Indexing.
>
>                      13.08.2004 06:45
>
>
>                      Please respond to        Category:
>      |-------------------------|
>
>                      "Lucene Users                            | ( )
>Action needed       |
>                      List"                                    | ( )
>Decision needed     |
>                                                               | ( )
>General Information |
>
|-------------------------|
>
>
>
>
>
>
>
>
>
>Natarajan,
>
>MultiSeacher - it is well, but this a way have pickles.
>
>Example, but it is not sample:
>
>public Query combine(Query[] queries) throws IOException {
>     if (expandedQueries.length < 2) {
>         return queries[0];
>     }
>     Query[] combined = new Query[2];
>     combined[0] = new BooleanQuery();
>     BooleanQuery.setMaxClauseCount(10000);
>     for (int i = 0; i < queries.length; i++) {
>         combined[1] = queries[i];
>         if (queries[i] instanceof BooleanQuery ||
>             queries[i] instanceof MultiTermQuery ||
>             queries[i] instanceof PrefixQuery ||
>             queries[i] instanceof RangeQuery) {
>             combined[0] = Query.mergeBooleanQueries(combined);
>         } else if (queries[i] instanceof PhraseQuery) {
>             Term[] queryTerms =
>((PhraseQuery)queries[i]).getTerms();
>             for (int j = 0; j < queryTerms.length; j++) {
>                 TermQuery q = new TermQuery(queryTerms[j]);
>                 ((BooleanQuery)combined[0]).add(q, true, false);
>             }
>         } else ((BooleanQuery)combined[0]).add(queries[i], true,
>false);
>     }
>     return combined[0];
>}
>
>...
>     Searcher[] searchers = new IndexSearcher[indexName.length];
>     for(int i=0;i<indexName.length;i++) {
>         searchers[i] = new IndexSearcher(indexName[i]);
>     }
>     MultiSearcher multiSearcher=new MultiSearcher(searchers);
>     QueryParser qp = new QueryParser(FIELD_CONTENTS, analyzer);
>     query = QueryParser.parse(queryString, FIELD_CONTENTS,
>analyzer);
>     hits = multiSearcher.search(query);
>     IndexReader reader[] = new IndexReader[indexName.length];
>     Query[] expandedQueries=new Query[indexName.length];
>     for(int i=0;i<indexName.length;i++){
>         expandedQueries[i]=query.rewrite(reader[i]);
>     }
>     query=combine(expandedQueries);
>...
>
>Best regards,
>Vladimir.
>
>
>
>
>
>On Thu, 12 Aug 2004 20:51:13 +0530
>  "Natarajan.T" <natarajant@crimsonlogic.co.in> wrote:
>>Thanks for your response.
>>Ok I can understand the concept ..... if you have any sample code pls
>>sent it to me.
>>
>>You have any idea about Parallel Searcher pls share to me.
>>
>>-----Original Message-----
>>From: Terence Lai [mailto:tlai@trekspace.com]
>>Sent: Thursday, August 12, 2004 8:40 PM
>>To: Lucene Users List
>>Subject: RE: MultiSearcher to Indexing.
>>
>>This is how I do it:
>>
>>IndexSearcher[] is = new IndexSearcher[2];
>>
>>is[0] = new IndexSearcher(IndexDir1); // first index folder
>>is[1] = new IndexSearcher(IndexDir2); // second index folder
>>
>>MultiSearcher searcher = new MultiSearcher(is);
>>
>>searcher.search(query);
>>
>>I think that the MulitSearcher is only doing sequential search.
>>Alternately, you can use ParallelMultiSearcher which allows you to do
>>the search in parallel.
>>
>>Hope this helps,
>>Terence
>>
>>
>>
>>> FYI
>>>
>>> I have an Indexing files in different folders, in this time how can
>>>I
>>> doing  the Searching process using MultiSearcher.
>>>
>>> Thanks,
>>> Natarajan.
>>>
>>>
>>>
>>>
>>>
>>>
>>
>>
>>
>>
>>----------------------------------------------------------
>>Get your free email account from http://www.trekspace.com
>>          Your Internet Virtual Desktop!
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
>>For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>>
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
>>For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
>For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>
>
>
>
>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
>For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>


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







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


Mime
View raw message