Re: unexpected behavior with certain words "i"

Tech Tip: Click here to run a free scan for Windows Errors and optimize PC performance

Thorsten wrote on Mon, 9 Jul 2007 02:38:02 -0700:

Hello,
I have started to test fulltext indexing. I have run into a problem that I
don't know how to solve.
I have this data in one fulltext-column "Landsorganisationen i Sverige".
When I perform this search I get 1 hit:
select top 100 * from CA_CATALOG_New_SEARCH where
CONTAINS(FWD_ALL,'"landsorganisationen"') AND
CONTAINS(FWD_ALL,'"sverige"')

But, when I do this search I get 0 hits (should be 1 hit):
select top 100 * from CA_CATALOG_New_SEARCH where
CONTAINS(FWD_ALL,'"landsorganisationen"') AND
CONTAINS(FWD_ALL,'"sverige"') AND CONTAINS(FWD_ALL,'"i"')

There is something with the word "i" that confuses the search engine. I
suspect that it is considered a special word and not indexed. But since I
want a hit, it seems like I need to know all words that won't work and
exclude them from searches. But it sounds like unnessecery work. It must
be some better way to work around this.
And if I need to know which words, were is the list?

Sincerely,
Thorsten

There is a list of words held in the noise.<languagecode> files (eg.
noise.enu for Neutral English) which are not indexed. You can remove all of
these (but leave a single line with a space on it), and then repopulate your
index. Just remember to edit the correct file related to your database
locale.

The other alternative is to parse the noise word list and create a routine
to strip out these ignored words when searching - the benefit of this is
that you keep the keep the index catalog small and strip out common words
from the searches. However, it might be more work than the simple clearing
of the noise file.

Personally, I cleared all my noise word files - the performance hit was
negligible for the tests I performed on my FT searching before and after.

Dan


.

Relevant Pages

  • My trip so far
    ... Monday I decided to play tourist and head down to the strip with my ... I then hopped on the Deuce and got the front seat in the ... me hit his royal and took the progressive of $1600 +, ... the ElCo playing roulette again and cashed out there up $60. ...
    (alt.vacation.las-vegas)
  • Re: Lio a hit
    ... It's interesting what a gulf there is between this level of "hit" and the ... or has the "perverse" tone of Lio seem to have been toned ... It's still a very clever and funny strip and I ... what works and what doesn't is all part of the growing process of any new ...
    (rec.arts.comics.strips)
  • Re: at Main Street
    ... I stayed up all night waiting for Cam's call. ... We hit the Strip today. ... Need coffee. ...
    (alt.vacation.las-vegas)
  • Re: Deathful Deer (Mark Trail, 9 February)
    ... strip except that Bucky the deer's been shot, ... if he's taken a hit in the head rather than the antler as Saturday ...
    (rec.arts.comics.strips)
  • Re: FBOW 19 Sep 06: Becky was right!
    ... :Renee wrote: ... :> April and Eva seem awfully mean in today's strip. ... :People who hit it big and then stop hanging with their old friends ...
    (rec.arts.comics.strips)