: I was was just a bit surprised to see that I had to go from objects to a
: string-rep.
Like i said you don't *have* do ... if you use Solr in an embeeded
context, or change your code to be a Solr plugin then you you can use
Query obecjts as is ... but when you send the data over the wire it has to
be serialized in some way.
: I did indeed also find out that some information was lost in that process,
: for instance setMinimumNumberShouldMatch on BooleanQuery. Also,
: some empty clauses (which was silently accepted by lucene) rendered
: as "()" which did not seem to be parsable on the solr-side. In other words,
: nothing I couln't workaround.
...those are all actually fundemental issue with the low level Lucene
query classes and the default Lucene QueryParser ... not really a Solr
issue.
: Good to know about the extension points though.
BEGIN: shameless plug...
And if you'd like to learn more about Solr's extension points, you should
sign up for ApacheCon and come to "Apache Solr: Beyond the Box" ...
http://us.apachecon.com/c/acus2008/sessions/10
"Apache Solr is an HTTP based enterprise search server built on top of the
Lucene Java search library. In this session we will look at Solr's
internal Java APIs and discuss how to write various types of plugins for
customizing it's behavior -- as well as some real world examples of "When"
and "Why" it makes sense to do so."
END
(And now i feel dirty)
-Hoss
