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: MaxFieldLength
Date Tue, 13 Jul 2010 16:42:44 GMT
Yep, that's pretty much it..

Erick

On Tue, Jul 13, 2010 at 12:47 AM, manjula wijewickrema
<manjula53@gmail.com>wrote:

> Ok Erick, answer is there. If there is no any document exceeds the default
> maxfieldlength, then no any document will be truncated although we increase
> the no. of documents in the index. A'm I correct? Thanx for your
> commitment.
>
> Manjula.
>
> On Tue, Jul 13, 2010 at 3:57 AM, Erick Erickson <erickerickson@gmail.com
> >wrote:
>
> > I'm not sure I understand your question. The number of documents
> > has no bearing on the field length of each, which is what the
> > max field length is all about. You can change the value here
> > by calling Indexwriter.setMaxFieldLength to something shorter
> > than the default.
> >
> > So no, if no document exceeds the default (Terms, not characters),
> > no document will be truncated.
> >
> > The 10,000 limit also has no bearing on how much space indexing
> > a document takes as long as there are fewer then 10,000 terms. That
> > is, a document with 5,000 terms will take up just as much space
> > with any MaxfieldLength > 5,000.
> >
> > HTH
> > Erick
> >
> > On Mon, Jul 12, 2010 at 4:00 AM, manjula wijewickrema
> > <manjula53@gmail.com>wrote:
> >
> > > Hi,
> > >
> > > I have seen that, onece the field length of a document goes over a
> > certain
> > > limit (
> > >
> > >
> >
> http://lucene.apache.org/java/2_9_3/api/all/org/apache/lucene/index/IndexWriter.html#DEFAULT_MAX_FIELD_LENGTH
> > > gives
> > > it as 10,000 terms-default) Lucene truncates those documents. Is there
> > any
> > > possibility to truncate documents, if we increase the number of indexed
> > > documents (assume, there are no any individual documents which exceed
> the
> > > default MaxFieldLength of Lucene)?
> > >
> > > Thanx
> > > Manjula.
> > >
> >
>

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