Xapian discuss - Feb 2007 - Indexing specific data.(Help required)

Hello,
 Thank you for your last replies regarding postgres connection.
I will try that sometime but I am approaching bit different way.
I have created some csv out of the postgres database and applying indexing
on that csv.

But problem is that my data is in Devnagari script and I use UTF-8 encoding
for it's support.
By applying some scripting called souindics which extracts sound code out of
the word and
store it with in English letters.
 So, I have to process first with above step and then with xapian
php-binding to index and search .
But in this process my original document gets besides and result appears as
my sound code instead.

So, is there any thing by which I can maintain the reference to original
document?.
Can it be possible to index only specific column of the csv?
Thank you

Regards

-- 
              (((())))
              (@ @)
                 (_)
+----oOO------------Ooo----------+
|      Gupteshwar D Joshi            |
|                                              |
+------------------------------------+
              |___|___|
                 | | | |
              ooO Ooo

On Wed, Feb 21, 2007 at 04:06:30PM +0530, Gupteshwar Joshi
wrote:
> But problem is that my data is in Devnagari script and I use UTF-8 encoding
> for it's support.
> By applying some scripting called souindics which extracts sound code
> out of the word and store it with in English letters.
> So, I have to process first with above step and then with xapian
> php-binding to index and search .
> But in this process my original document gets besides and result appears as
> my sound code instead.
Note that there's no reason why you can't store the UTF-8 Devnagari
script version in the document data, but generate terms from the
anglicised version.

It occurs to me that this "sounindics" is a term normalisation
procedure,
so it's a lot like a stemming algorithm in many ways.  I can't find
any information on Google about it though - "sounindics" has no
matches
and "sounindic" only finds a passing mention in someone's CV (a
reference to using Xapian in fact!)

But perhaps this algorithm should be wrapped as a Xapian::Stem class,
which would make it very easy to index and query Devnagari script
in this way.  Do you have a reference for it?
> So, is there any thing by which I can maintain the reference to original
> document?.
An alternative approach is to store a unique id for the postgres
database record in the Xapian document data.
> Can it be possible to index only specific column of the csv?
Erm, of course.  Just only generate terms from that column!

Cheers,
    Olly