Xapian discuss - May 2007 - Multiple matches in the same file

Hi,

I am new to xapian (and rather satisfied, by
the way. Congrats to the developers). However, I would be
happier if it were able to return multiple matches for
the same document in a coherent fashion. I mean the
obvious: not just returning the document once but the whole
list of document+matches_in_the_document for each "match".

I know this does not make sense for many queries (AND's
specifically), but it would be more than useful for NEAR
queries and ORs, in my opinion.

I may be wrong, though, but IIRC, this is not possible
**at present**, but could be implemented. I may even
venture to try to (and re-get my C++), but I need to
be certain that it does not exist.

The interest is, for example, for literary studies and
for large texts (think of law codes, which are full of
similar entries which may be all of interest).

Thanks,

Pedro.


Pedro Fortuny Ayuso
pfortuny@gmail.com      http://pfortuny.sdf-eu.org
C/Capuchinos 14, 1-S. 47006 Valladolid. SPAIN

On Sat, May 05, 2007 at 12:50:04PM +0200, Pedro Fortuny Ayuso
wrote:
> I am new to xapian (and rather satisfied, by
> the way. Congrats to the developers). However, I would be
> happier if it were able to return multiple matches for
> the same document in a coherent fashion. I mean the
> obvious: not just returning the document once but the whole
> list of document+matches_in_the_document for each "match".
I don't think I really understand what you're asking for.  What may be
"the obvious" to you isn't to me!

In Xapian a "document" is the unit that is retrieved - it needn't
be
the same as what you might naturally consider a document in the system
being indexed.  If you want to retrieve with a finer granularity, then
you can just define your "Xapian document" to be that granularity. 
You
can collapse matches within the same larger unit together if you wish.

If that isn't what you mean, can you elaborate?

Cheers,
    Olly