Re: [htdig] How to find all doc which content "word and word2"?

[Thierry Herve]

Geoff Hutchison wrote:

At 11:36 AM +0200 4/27/00, Thierry Herve wrote:
>Actually, I index my site with htdig. If I make search with keyword,
>for example, "TMN and CR", htdig found all doc which content TMN and
>CR words both but it also find all docs which content TMN word only
>and all docs which content CR word only.

I think you have set match_method to "any" or "or" because the
default will include all words and will do what you want. One note:
if you have not set minimum_word_length, then CR will be ignored
because it is smaller than the default (3).
 

In fact, I don't use htdig in standard use (from web server) because I have not roght access to install it. Then I use it from command line. I use the following script :
 

#! /bin/sh -f

dirHtSearch=/home/herve/tmp/test_htdig/htdig_3.1.2/cgi-bin
dirNetscape=/soft_hp/NETSCAPE47
searchResultFile=/home/herve/tmp/search_result.html

pageNumber=$1
shift
searchWords=""
inputLine="$*"
for token in $inputLine
do
  searchWords="${searchWords}+${token}"
done

export REQUEST_METHOD='GET'
export QUERY_STRING='config=&restrict=&exclude=&method=boolean&format=builtin-long&sort=score&words='${searchWords}'&page='${pageNumber}

echo $QUERY_STRING

$dirHtSearch/htsearch > $searchResultFile
$dirNetscape/netscape -remote 'openURL( file:'${searchResultFile}' )'

but, this is no important to me for the moment.

here, I give the CGI URL (I don't know exactly his name). The match_method I used is boolean.

config=&restrict=&exclude=&method=boolean&format=builtin-long&sort=score&words=+TMN+and+gestion&page=1

Thank's for minimum_word_length info. I re-build my data base.

 
--
-Geoff Hutchison
Williams Students Online
http://wso.williams.edu/

-- 

---

[EMAIL PROTECTED]
Voice : +33 (0)1 46 52 33 49
Fax     : +33 (0)1 46 52 42 80