Is it possible to modify a stored field but not indexed? for example, if I
have a field like this:
new Field("address", address, Field.Store.YES, Field.Index.N)
and I want to modify it like this:
hits.doc(i).getField("address").set("11 Diana Street");
Is it possible?

but if you remove it and add it again, you'll need to re-index it again.
don't you? When you do re-index, you'll have to close the reader, which
would pause the search. Any better way of doint it?

Samstag 08 Juli 2006 00:03, dan2000 wrote:
When you do re-index, you'll have to close the reader, which
would pause the search. Any better way of doint it?
Try using IndexModifier (added in Lucene 1.9).
Regards
Daniel

Thanks a lot Doron.

I thought I had to close all readers and searchers in order to new
IndexWriter, otherwise I keep getting Lock timeout exception in a
multithreaded environment. In my case, modifying/adding/deleting only
happens occasionally, so I don't have a IndexWritter that is open all the
time. In order to avoid lock timeout exception, I had to put a real lock in
my search method, and a write lock in my add/modify method. I know this is
not efficient enough since all the readers/searcher are re-opened each time
adding/modifing happens.

From what you said, I'm thinking of switching to IndexModifier. Before that,
I have a few more questions:
1. without using an exclusive lock, is there anyway to new a IndexModifier
or IndexWritter while IndexReaders are serving search queries? or do I have
to new a IndexWriiter when system starts and keep it open all the time?
2. When adding a new document, and calling optimize(), do I have to close
the readers/searchers?
3. In my case, updating non-idexed fields happens a lot, can you explain the
side data store in more detail? I'm not sure with this approch. Do you mean
creating a seperate database and stores lucene's ids?

Thanks a lot Doron. I'm gonna give it a try now. The problem I've had before
was that I set my writer to null right after close it. That's why I got lock
timeout exception when i try to create a the writer again. Guess I just need
to close it, and re-open it would avoid the locking problems then.

Thanks.

Here is the simplified code that causes problem (Lock obtain timed out).
MyIndexer is used for indexing and searching. IndexTest starts 5 threads for
indexing and 100 threads for searching.

MyIndexer.java
public class MyIndexer
{
File m_IndexFile;
IndexReader m_IndexReader;
Directory m_Directory;
Searcher m_Searcher;
Analyzer m_Analyzer;
ReentrantReadWriteLock lock = new ReentrantReadWriteLock();

public MyIndexer(File a_IndexFile) throws IException
{
m_IndexFile = a_IndexFile;
m_Analyzer = new StandardAnalyzer();
m_Directory = FSDirectory.getDirectory(m_IndexFile, false);
m_IndexReader = IndexReader.open(m_Directory);
m_Searcher = new IndexSearcher(m_IndexReader);
}

public void addAndIndex(String a_Name) throws IException
{
// lock.writeLock().lock();
try
{
IndexWriter indexWriter = new IndexWriter(m_Directory,
m_Analyzer,
false);
Document doc = new Document();
doc.add(new Field("name", a_Name, Field.Store.YES,
Field.Index.TKENIZED));
indexWriter.addDocument(doc);
indexWriter.optimize();
indexWriter.close();
}
finally
{
// lock.writeLock().unlock();
}
}

public Hits search(String a_Query)
{
try
{
// lock.readLock().lock();
Query query = new QueryParser("name",
m_Analyzer).parse(a_Query);
Hits hits = m_Searcher.search(query);
return hits;
}
catch (Exception e)
{
e.printStackTrace();
}
finally
{
// lock.readLock().unlock();
}
return null;
}
}

IndexTest.java
public class IndexTest extends Thread
{
private boolean m_DoIndex = true;
static MyIndexer m_IndexEntity;

public IndexTest(boolean a_DoIndex)
{
m_DoIndex = a_DoIndex;
}

public static void main(String[] args) throws Exception
{
m_IndexEntity = new MyIndexer(new File(
"C:/mytest/myindex"));

for (int i = 0; i < 5; i++)
{
new IndexTest(true).start();
}

for (int i = 0; i < 100; i++)
{
new IndexTest(false).start();
}

}

@
public void run()
{
for (int i = 0; i < 50; i++)
{
try
{
if (m_DoIndex)
{
String name = "Nick" + i;

System.out.println("Add and Indexing");
m_IndexEntity.addAndIndex(name);
Thread.sleep(100);
}
else
{
System.out.println("Searching");
m_IndexEntity.search("name:Nick" + i);
Thread.sleep(20);
}
}
catch (Exception e)
{
e.printStackTrace();
}
}
}
}

Thanks for your advice Doron. I've tried changing to one indexing thread
(instead of 5) but still get the same problem. can't figure out why this
happens.

I did clean everything but still getting the same problem. I'm using lucene
2.0. Do you get the same problem on your machine?

can't access the file:
Forbidden
Remote Host: [62.172.205.164]
You do not have permission to access
Data files must be stored on the same site they are linked from.
Thank you for using 20m.com

Thanks Doron. The site you provided doesn't support firefox, that's why had
problem with downloading.

Your code works fine and I've just noticed I didn't change the create
parameter to false when I've leaned the index directory. Sorry for my
mistake.

Thanks a lot for your help Doron. You advice helped me a lot.