Xapian discuss - Jun 2007 - Z's and other prefixes

Hi,

I've been playing around with xapian, and I have a couple of questions.

I know that QueryParser and TermGenerator are matched in the way they come up
with terms, based on stemming and all that. They prefix "Z" to the
start of a
stemmed word.

I'm trying to come up with a solution to index and retrieve exact documents
based on their names. And I was thinking that I may use my own prefix to make
sure that page comes up as 100% (or at least on the first page).

Does anything else use prefixes? I haven't played around with the DateRange
things, but do they use specific prefixes?

I don't want to conflict with an existing prefix. I also may use a
multi-character prefix to avoid subsets and other things.

Cheers.

On Fri, Jun 15, 2007 at 01:10:16AM +0000, David wrote:
> Does anything else use prefixes? I haven't played around with the
DateRange
> things, but do they use specific prefixes?
The DateValueRangeProcessor approach uses values rather than terms, so
no term prefixes are involved.  Omega can use a term-based approach to
date ranges, and that uses D, M, and Y, but now also offers a
value-based approach.  I suspect that the term-based approach will be
phased out eventually (but currently I've not benchmarked the two and
I'd like to check that the new approach works better before removing
the old one!)

Prefixes can be used for probabilistic fields (e.g. to allow searching
within the title of a document), and for boolean filters (e.g. to allow
searches to be restricted to a particular document type).

This documents the prefixes which are standardly used:

http://svn.xapian.org/trunk/xapian-applications/omega/docs/termprefixes.txt?view=markup

Cheers,
    Olly