Mailing-List: contact general-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: general@lucene.apache.org
Received-SPF: pass (asf.osuosl.org: domain of fidde.andersson@gmail.com
 designates 66.249.82.200 as permitted sender)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
        s=beta; d=gmail.com;
        h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type;
        b=nAHU0vASOVCxSq1aNvc0IMhnL97Uj9XK1UshmwxLbYguZRZb2Kshjy/Ng/4ZDvw5OS4Zm4XrE5Umjp2Hq3dcBVd3Dnp6OOZk6qQTvAWwt83tqKLMaOlLpavEgM9kX9RWfU1USPXM1YicF3D2+c4g93wSCOGaOJl6cqszxJVEfYM=
Message-ID: <b4c6e18805092900452fe350c3@mail.gmail.com>
Date: Thu, 29 Sep 2005 09:45:40 +0200
From: Fredrik Andersson <fidde.andersson@gmail.com>
Reply-To: Fredrik Andersson <fidde.andersson@gmail.com>
To: general@lucene.apache.org
Subject: Reading back binary fields
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_Part_1814_19110431.1127979940121"

------=_Part_1814_19110431.1127979940121
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Hey Gang!

I'm having some problems when modifying an existing index, adding a binary
field to each document. Or more specifically, I have a problem reading back
that field. I'm using the IndexModifier from the trunk, and I am positive
that the binary field gets written down, since the field name shows up in i=
n
Luke.

The index modification is nothing out of the ordinary.
-----------------
int i =3D some integer
Document doc =3D indexReader.document(i);
indexModifier.delete(i);
Field tvField =3D new Field(TV_FIELD_NAME,aByteArray,Field.Store.YES);
doc.add(tvField);
indexModifer.addDocument(doc);
indexModifier.optimize();
indexModifier.close();
-----------------

However, when I re-read that document at a later point in time, and iterate
the doc.fields() enumeration, the TV_FIELD_NAME field does not show up. Luk=
e
shows that field, though - what's the trick?

On a side-note, I have modified the IndexModifier to use write/delete
buffers, so that when some defined buffer size overflows, the buffers
flushes down their stuff to the internal indexReader and indexWriter in
delete-write order. It increased the speed tremendously compared to the old
IndexModifier, which doesn't use any form of internal buffering. If this is
something you want comitted to the Lucene trunk, give me a shout and I'll
clean it up.

Thanks,
Fredrik

------=_Part_1814_19110431.1127979940121--