[Xapian-discuss] Stopper Problems

[Olly Betts]

On Wed, Mar 07, 2007 at 04:08:57PM +0000, Colin Bell wrote:
> I have made a stopper which I use for indexing (and retrieval) which  
> doesn't seem to stop much. (I've pasted a trimmed example Below) .  
> Loads of the words in the stopper keep coming through.
> 
> I use it as follows:

The code looks plausible to me.  It's not a complete program, so I
couldn't easily try running it.

Do you have some example queries where stop words aren't removed?

Note that the QueryParser doesn't stop words in phrases, or with "+"
in front of them.  There are a few other cases too.  Essentially it
expects search time stop word removals rather than index time, but the
behaviour with index time stop word removal is to fail to match
phrases and "+" terms which contain/are stop words, which isn't too
unreasonable.  If we can decide on a bettere way to handle these cases
when the word wasn't indexed, it wouldn't be hard to change this.

Cheers,
    Olly