lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Lucas F. A. Teixeira" <lucas.teixe...@accurate.com.br>
Subject Re: Lucene CFS naming significance
Date Fri, 27 Jun 2008 11:12:33 GMT
Folks,
Could anyone tell me the significance of the naming of the cfs files in the
luceneindex e.g.  _1pp.cfs, _2kk.cfs etc.

 > Just names that won`t repeat in the same folder.

I have observed many differently named files being created temporarily while
the index is being built, but the same set of named files are in place once
the index has finished building.

 > That not true. Unless you always have the same "data" every time you 
build the index, and if you build it every time from the beggining (not 
rewriting the docs)

Can I rely on the index files always having the same names once index
buildng is complete? 

 > No. That names will be rotating through your indexing.

This would greatly simplify my auto ftp'ing of the
files up to a web server via an SSIS package.

 > Why don`t you just ftp the hole index dir? The directory is actually 
the "index" by itself, and not its content.

[]s,


Lucas Frare A. Teixeira
lucas.teixeira@accurate.com.br <mailto:lucas.teixeira@accurate.com.br>
Tel: +55 11 3660.1622 - R3018



mick l escreveu:
> Folks,
> Could anyone tell me the significance of the naming of the cfs files in the
> luceneindex e.g.  _1pp.cfs, _2kk.cfs etc.
> I have observed many differently named files being created temporarily while
> the index is being built, but the same set of named files are in place once
> the index has finished building.
> Can I rely on the index files always having the same names once index
> buildng is complete? This would greatly simplify my auto ftp'ing of the
> files up to a web server via an SSIS package.
> Thanks
> M
>   

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