Shai,
Thanks for your response. There isn't any specific options I am setting
and am leaving everything to defaults:
IndexWriter indexWriter = new
IndexWriter(FSDirectory.getDirectory(stemIndexPath), new
WhitespaceAnalyzer());
indexWriter.setMergeFactor(10);
indexWriter.setInfoStream(System.out);
Documents are added using:
Document doc = new Document();
Field fld1 = new Field("fld1", "value1", Field.Store.NO,
Field.Index.UN_TOKENIZED);
doc.add(fld1);
Field fld2 = new Field("fld2", "value2", Field.Store.YES,
Field.Index.NO);
doc.add(fld2);
indexWriter.addDocument(doc);
I added the following:
indexWriter.setMaxBufferedDocs(10);
No difference - it continues to create one document in each RAM segment
before the first merge.
-venkat
-----Original Message-----
From: Shai Erera [mailto:ser...@gmail.com]
Sent: Saturday, July 25, 2009 10:55 PM
To: java-user@lucene.apache.org
Subject: Re: Number of documents in each segment before a merge occurs
How do you set up IndexWriter (can you paste the code here)?
setMergeFactor
just controls how often merges will occur. But there are other settings,
like setRAMBufferSizeMB and setMaxBufferedDocs that control how often to
create a segment. If for example you set maxBufferedDocs to 1, or
RAMBufferSizeMB to a small value, such that each document fills the
entire
buffer, you'll run into the scenario you mentioned.
If that's not the case, then it'd be good if you can paste here the
indexing
code including the IndexWriter setup, docs indexing etc.
Shai
On Sun, Jul 26, 2009 at 8:25 AM, Venkat Rangan <
venkat.ran...@clearwellsystems.com> wrote:
> Hi,
>
>
>
> We are writing to an index using IndexWriter with setMergeFactor() set
> to 10. After inserting several documents, a merge occurs, but our
> observation is that each segment only has a single document. Is there
a
> way to get more than a single document in each segment before they are
> picked up for a merge?
>
>
>
> The following is the merge debug line when setInfoStream is set to
> stdout stream.
>
>
>
> 2009-07-26 07:07:09,202 INFO [STDOUT] (main:) merging segments _ram_0
> (1 docs) _ram_1 (1 docs) _ram_2 (1 docs) _ram_3 (1 docs) _ram_4 (1
docs)
> _ram_5 (1 docs) _ram_6 (1 docs) _ram_7 (1 docs) _ram_8 (1 docs) _ram_9
> (1 docs) into _0 (10 docs)
>
>
>
> Thanks,
>
>
>
> Venkat Rangan
>
> Clearwell Systems Inc.
>
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-user-h...@lucene.apache.org
