[Xapian-discuss] MSet order

[Pierre-Alain Moellic]

You've got it ! It was a problem with get_maxpart(), my estimation of the
upper bound wasn't correct.
Thank you Richard for this help.

I was wondering if it was possible to sort the results by relevance but in a
reverse order (smallest weight as first result, since - contrary to the
cosine similarity measure - classical distance give 0 for similar docs)

pa.

2011/3/8 Richard Boulton <richard at tartarus.org>

> Did you implement the get_maxpart() and get_maxextra() methods
> correctly?  It sounds like those are returning incorrect values, which
> would cause optimisations to fire early, and would produce effects
> such as those you describe.
>
> If you're able to post the source code somewhere (eg, in a site like
> pastebin.com), we might be able to spot the error.
>
> --
> Celestial Navigation Limited, incorporated in England & Wales
> (registration number 06978117), registered office address: 58
> Kingsway, Duxford, Cambridgeshire, CB224QN, UK.
>