Hi,
I am playing with MemoryIndex for a situation in which I have a large
number of small, ephemeral documents that I need to fire queries
at. It
appears to be at least 5x faster than RAMDirectory for my usage, which
is large enough to be interesting.
However MemoryIndex does not seem to support per field boosts,
which is
something we need. I presume its not a terribly difficult thing to
extend MemoryIndex to support per field boosts, but I am looking for
pointers from any lucene/MemoryIndex wizards on exactly how to attack
this.
Any guidance would be much appreciated.
thanks,
gt
Thanks for the reminder, I've now added support for per field boost
to the SVN trunk of MemoryIndex. This comes in the form of the new
method
public void addField(String fieldName, TokenStream stream, float boost)
To get it to compile against the latest stable release (2.0.0) you'll
need to simply remove the following method:
public Document document(int n, FieldSelector fieldSelector)
(The FieldSelector interface has nothing to do with boosting, and was
added after 2.0.0. shipped.)
Wolfgang.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
