Index file compressing (was: RE: Lucene.Net project involvement)

[George Aroush <ge...@aroush.net>]

Hi James,

I believe you are seeing such a high compression ration because you have
stored the indexed data into the Lucene index.

No, Lucene doesn't have an API to take an index as ZIP file.  The class
SharpZipLibAdapter is an adapter class which you must use when you are
indexing -- it will compress your indexed data as they are stored into the
Lucene index.

You must either re-index your data using SharpZipLibAdapter to tell Lucene
to compress your stored field data, or if you can't do this, your other
option is to use Windows's built in folder compression (on the folder where
the index is located.)

-- George Aroush
 

-----Original Message-----
From: Shaw, James [mailto:James_Shaw@intuit.com] 
Sent: Monday, March 26, 2007 8:26 PM
To: lucene-net-dev@incubator.apache.org
Subject: RE: Lucene.Net project involvement

Hi, guys,
I want to reduce Lucene index deployed on our customers' desktops to as
small as possible (for patching purpose), and I find that the index size
shrinks by 75% if I zip it.  Can I have Lucene use this zipped index (either
directly or indirectly) to append to it, initiate a search, etc?

I saw a RAMDirectory class, but I don't see methods to add file streams to
it (say I use SharpZipLib to uncompress the zipped index).  Does using
RAMDirectory have any drawback?