[Xapian-discuss] New QueryParser!

[Tim Brody]

Eric B. Ridge wrote:

> On Jan 28, 2005, at 5:30 AM, James Aylett wrote:
> 
>> -1 on this - that's really something that lots of people /won't/ want.
> 
> Why?  For those of us that don't use Omega having something builtin that 
> we can use to index text documents into terms that match what the 
> QueryParser parses makes a lot of sense to me.
> 
> Olly pointed out many moons ago, and I discovered on my own in dealing 
> with the JNI wrappers that it's super important that your text and query 
> parsers parse text exactly the same.  Otherwise you might index terms 
> you cant query for.

More importantly you might query for terms that you never indexed ...

While an "out of the box" parser might be useful to have, the query 
parser can't behave in the same way as the document parser as it needs 
to treat some terms as special. What's more important is that 
stemming/stopwords (language issues) are kept in sync on both sides, 
which probably goes back to more tightly associating these with the 
database.

Regards,
Tim.
(Still waiting for an updated Perl lib :-)