The noise word file used is the one corresponding to the language you
specified during the full-text index creation.
In the full-text indexing Wizard "Select Table Columns" step, you define a
"Language for Word Breaker".
For example, we are using "French (France)" in our application, and it works
fine.
I recommand you not to empty the noise words files, you'll loose some great
capabilities of full-text search.
For example, if someone search for a letter/number or a common word ('the',
'or'...), those words should be ignored, otherwise you'll get plenty of
"noise" in your results.
Even if it is minimal, the noise words file should never be empty.
Perhaps, should you try to delete all the full-text indexes and re-create
them from scratch?
Benoit Hediard
www.benorama.com
-----Message d'origine-----
De : Andy Ewings [mailto:[EMAIL PROTECTED]]
Envoyé : lundi 14 octobre 2002 17:18
À : CF-Talk
Objet : RE: Full Text Indexs
this doesn't appear to be the case here. "After eight" still throws an
error as "after" is reserved. eight however is ok. I have checked the
noise.eng file and it is empty. Is there any weay of telling which noise
file is being used? - I assume it's the English one
-----Original Message-----
From: Benoit Hediard [mailto:[EMAIL PROTECTED]]
Sent: 11 October 2002 18:08
To: CF-Talk
Subject: RE: Full Text Indexs
It should only throw an error if all the search words are defined in the
noise words file.
For example :
"The" will throw an error,
"The foo" will return results based on "foo" search.
In our application, we have customized the noise word file and put each full
text SQL search in <cftry> blocks.
If an error is thrown we create an empty recordset.
For example :
<cftry>
<cfquery name="qSearch" datasource="#application.stConfig.DSN#">
select ID_NEWS as ID,
TITLE_NEWS as Title,
CREATEDBYID_NEWS as CreatedByID,
CREATEDBY_NEWS as CreatedByLogin,
CREATIONDATE_NEWS as CreationDate
from ITEM_NEWS
and (contains(TITLE_NEWS,'"#arguments.Text#"')
or
contains(DESCRIPTION_NEWS,'"#arguments.Text#"'))
</cfquery>
<cfcatch type="Database">
<cfset qSearch =
queryNew("ID,Title,CreatedByID,CreatedByLogin,CreationDate")>
</cfcatch>
</cftry>
Benoit Hediard
www.benorama.com
-----Message d'origine-----
De : Andy Ewings [mailto:[EMAIL PROTECTED]]
Envoyé : vendredi 11 octobre 2002 18:41
À : CF-Talk
Objet : RE: Full Text Indexs
Hi
We tried this but it didn't seem to work. We only changed the english one.
When we search on these key words it still throws an ODBC error. Any ideas?
Andy
-----Original Message-----
From: Benoit Hediard [mailto:[EMAIL PROTECTED]]
Sent: 11 October 2002 13:23
To: CF-Talk
Subject: RE: Full Text Indexs
If you are using SQL Server 2000, all the "noise words" are defined in a
text files, one per language, you can easily modify them (add/remove words)
:
"../Microsoft SQL Server/MSSQL/FTDATA/SQLServer/Config/noise.eng" for
english indexing
"../Microsoft SQL Server/MSSQL/FTDATA/SQLServer/Config/noise.fra" for french
indexing
.
Once modified, you'll have to rebuild your full-text indexes.
Benoit Hediard
www.benorama.com
-----Message d'origine-----
De : Andy Ewings [mailto:[EMAIL PROTECTED]]
Envoyé : vendredi 11 octobre 2002 14:00
À : CF-Talk
Objet : Full Text Indexs
Bit OT but thought I'd test the SQL gurus on here - I have built some full
text indexes and am trying to get it to ignore the "ignored" words if that
makes sense. For exaple if I type "after eight" into a search box it gives
an error as "after" is a reserved word. How to I get SQL to ignore these
words when performing my fulltext query?
Alternatively does anyone know where I can find a full list of these ignored
words?
Andy
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
Archives: http://www.houseoffusion.com/cf_lists/index.cfm?forumid=4
Subscription: http://www.houseoffusion.com/index.cfm?sidebar=lists&body=lists/cf_talk
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Structure your ColdFusion code with Fusebox. Get the official book at
http://www.fusionauthority.com/bkinfo.cfm
