Xapian discuss - Apr 2006 - how to access position list

Hello

I have created a file search engine. Along with each term indexed, I am
storing the line number of line on which it occurs( as the terms position).
I have done this in PHP.Now i want to access the position of terms of each
document in mset. How can i do it? I dont think Mset class has any function
which directly gives positions of each matching term of each doc in
MSet.Tomake more clear what i want, I will an example. Suppose theres
a doc "
story.txt". This is an extract from the doc


line 99:   Finally the prince rescued her from the clutches of the evil
King.
line 100: And they lived ever after..


query : happily

now "story.txt" will be one of the results coz its index
"happily" matched.
Now i want to display the line on which "happily"  appeared in the
original
doc. While indexing, along with each term i am also storing the line on
which it appears.( Thus pos(prince)=99, pos(rescued)=99 ,pos(they)=100,
pos(lived)=100 and so on). On seacrching i will obtain the MSet. Now along
with things like docid, rank percentage doc data, I also want to display the
line from original doc, on which this word "happily" appears( Like how
we
see the results in google where the line from the html page where the search
term appeared, is displayed). For that I need access to the position list of
term "matching" in doc "story.txt".( say "matchin"
appears on lines 12, 45
and 99) How do i obtain that? Plz help.

Thanks

Durga
doubtfire40008@gmail.com

On Mon, Apr 03, 2006 at 09:02:26PM +0530, durga bidaye wrote:
> On seacrching i will obtain the MSet. [...] I need access to the
> position list of term "matching" in doc "story.txt".
Okay - the MSet gives you an MSetIterator, which gives you the
Document.

Similarly, the original Query gives you a TermIterator for the terms
in the query (Xapian::Query.get_terms_begin()).

Using that you can find all the terms in the Query, and for each use
the TermIterator in Document (Xapian::Document.termlist_begin()),
combined with the skip_to() method on TermIterator (eg:
skip_to("happily")) to move to the right term. Then the TermIterator
has positionlist_begin() giving you a position iterator which gives
you the actual termpos (in your case, the line number).

You can also get to the position list directly from Xapian::Database
using positionlist_begin(doc_id, tname), which may prove shorter in
code terms.

J

-- 
/--------------------------------------------------------------------------\
  James Aylett                                                  xapian.org
  james@tartarus.org                               uncertaintydivision.org