I'm getting back into the spatial search stuff and wanted to get some
feedback before starting down any path...
I'm building an app, where I need R-tree like functionality -- that is
search for all items within some extent / that touch some extent. If
anyone has ideas for how this could map to tokens, that would be
great... but i don't know of anything yet.
Currently, I index an extent as "XMin XMax YMin YMax", then have a
custom filter that loads an in memory R-Tree for each Reader from the
FieldCache. This works great, but the memory requirements get big as
the number of documents gets large. Also it takes a while to warm up.
With the new flexible indexing stuff, would it be possible to natively
write an rtree to disk in the index process? Somehow using the lucene
doc id? Alternatively, I can look at writing the spatial index to
some other format (perhaps http://hatbox.sourceforge.net/) and then
try to keep lucene doc ids in sync with application ids.
Any thoughts on what may be possible?
It looks like Lucy has considered something similar:
http://www.lucidimagination.com/search/document/75ac07b7e2d6160d/pluggable_indexreader_was_real_time_updates
thanks
ryan
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-user-h...@lucene.apache.org
