I am not using DSpace just yet, but lucene in another application. This
is an extract to show how the index is pointed out in my case (a
directory on the same server as the application):
public static final String D_LUCENE_INDEX_DIR =
static final String DATASOURCE_NAME = "harvestdb";
private static final String LUCENE_DEFAULT =
String dir = System.getProperty(D_LUCENE_INDEX_DIR, LUCENE_DEFAULT);
File fDir = new File(dir);
if (!fDir.exists() || !fDir.isDirectory() || !fDir.canWrite())
throw new ServletException("some good message");
protected Directory indexDir;
indexDir = NIOFSDirectory.getDirectory(fDir);
We use Lucene as a black or dark grey box, so I cannot answer the other
two questions. However, I have noticed that performance is very
dependent on optimizing the index. In some cases (after heavy reindexing
e.g.) more than twice the "normal" disc space is needed during
Swedish National Heritage Board
>>> 2010-11-26 kl. 07:53, i meddelandet
<BAY147-w21DB25698E96C4CCCA65FDE1210@...>, skrev AMJAD USMAN
> hello everybody,
> i am a research student. my interest lies in information retrieval in
> digital repositories especially in document indexing and retrieval.
> My basic objective is to improve searching facility provided in
> repository systems especially in DSpace.
> As far i know Dspace uses Lucene for document indexing and searching,
> could not find any detailed documentation that help me understand:
> how Lucene creates Index?
> where it stores its index ?
> what technique it uses for searching documents?
> one of my colleague is working on "Implementation of multi-faceted
> in DSpace as like DBLP, Flamenco etc"
> can anybody help or guide in this regard...
> National University of Science and Technology, Pakistan