Setting apicp exactphrase parameters

WI-User · Post by **WI-User** » Tue Apr 17, 2001 4:58 pm

I want to search for the documents with the exact phrase. I have the following setting in the script newsearch:

<switch $sufs>
<case 0><apicp exactphrase 1>
<case 1><apicp suffixproc 1>
<apicp minwordlen 5>
<$x= "'" "s" "es"><apicp suffix $x>
<case 2><apicp suffixproc 1>
<apicp minwordlen 5>
</switch>

However when I search for say "Hello world" (without quotes) it seem to be fetching documents with noise in between. i.e. I get documents with strings "Hello there world" in it. However if I put "Hello world" within quotes it seem to be working. Is there anyway I can make "Hello world", without quotes, to match only those documents with exact phrase "Hello world"

Thanks

Post by **John** » Tue Apr 17, 2001 5:49 pm

No, although you could add quotes to what the user entered fairly easily.

The meaning of exactphrase is a little different than you think. Setting exactphrase will check that any noise words in a phrase are matched. For example if you search for "state of the art" then with the default settings Texis will look for "state" followed by any two words followed by "art".

WI-User · Post by **WI-User** » Wed Apr 18, 2001 1:22 pm

I am bit confused with the example you have given. Does it just look for only "any two" words between "state" and "art" or "any number of words" ?

I mean does it return documents containing "state in art", "state of art", "state in the complex art" etc..

If I search for "blood cancer" without quotes at "www.personalmd.com/search.shtml" with "exact phrase" as search criteria then I am getting documents with "Blood Test May Detect Cancer", "Blood Test Predicts Prostate Cancer". Is this correct ?

Thanks for your help.

Post by **John** » Wed Apr 18, 2001 1:46 pm

If you do not use double quotes then the words are not treated as a phrase, so the results you see for blood cancer are correct. If you want to search for a phrase you need to use double quotes, which could be added by the script if that option is chosen.

The example was searching for "state of the art" with the double quotes making it a phrase. It would not find any of your examples, but would find a document containing: State of Nebraska Art Festival.

r.marshall · Post by **r.marshall** » Thu May 24, 2001 4:22 am

I am a new user to the Texis product and have a questions regarding NOISE words and searching in general.

Question ...
I enter "About Morse" (without quotes) as my query and I expect to get at least 1 result back because About Morse is a top level menu item and appears in the <TITLE> tags of my page. I understand that "about" is a noise word and should be ignored and the search should take place with just "morse" (no quotes) as the query. As the website contains the word morse everywhere I expect a large number of hits .... but, my results tell me that no documents where found.

Any help would be appreciated.

Thanks,
Rob.

Post by **Kai** » Thu May 24, 2001 10:34 am

Look at the HTML source of the results page in your browser: there may be error messages in comments; what are they? Is this a standard Webinator search? If not, what is the exact SQL query, and what indexes exist on the table?

r.marshall · Post by **r.marshall** » Fri May 25, 2001 12:04 pm

There are no error messages in the html. After investigation (and some support) it turns out that the url I was creating to call the search script wasn't encoding the query,
e.q.
About Morse was http://......&query="About Morse".
It should be http://......&query="About+Morse"

cheers,

Rob Marshall