| Problem with stop words [message #272371] |
Thu, 04 October 2007 14:56  |
amazing
Messages: 46
Registered: September 2007
Location: Venezuela
|
Member |
|
|
Hi. I create my own stop list with the preference BASIC_STOPLIST that have my personalized stop words. For some contains query operators, when return a stop word, it do not appear in the resulting query, or highlighted. Doesn't it..?
Here is the problem. I have an 'Y' stop word. And when I use for example
...contains(text, '$Y', 1) > 0 ...
the resulting query or highlighted comes with that word and it also happen with the others stop words. So what am i doing wrong...???
I'm dealing with Oracle 10g patch 10.2.0.3 and i know that it makes some problems. Am I dealing with this kind of problem?
|
|
|
|
|
|
|
|
|
|
| Re: Problem with stop words [message #272588 is a reply to message #272580] |
Fri, 05 October 2007 10:57  |
amazing
Messages: 46
Registered: September 2007
Location: Venezuela
|
Member |
|
|
If I query with a text that includes a stop word, and use an operator that when return a stop word, do not include this stop word, the resulting query also includes a stop word. I' am developing an spanish solution: For example, a stop word is 'Y'. If i query by 'BANCOS Y TIENDAS', the resulting query is like:
'BANCOS Y TIENDAS'
'... Y ...'
'Y...Y...'
|
|
|
|
| Re: Problem with stop words [message #272648 is a reply to message #272588] |
Fri, 05 October 2007 15:24 |
yumyumyeh
Messages: 8
Registered: October 2007
|
Junior Member |
|
|
It looks like your index has not really been updated with your stopword.
Example:
create table test_index(search_text CLOB);
INSERT INTO test_index VALUES ('on?demand');
INSERT INTO test_index VALUES ('on demand');
INSERT INTO test_index VALUES ('on demand?');
BEGIN
CTX_DDL.CREATE_PREFERENCE ('test_lexer', 'BASIC_LEXER');
CTX_DDL.SET_ATTRIBUTE ('test_lexer', 'PRINTJOINS', '?');
END;
CREATE INDEX i_tk_test_index ON test_index (search_text) INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS ('LEXER test_lexer STOPLIST CTXSYS.DEFAULT_STOPLIST');
SELECT token_text FROM dr$i_tk_test_index$i
SELECT * FROM test_index WHERE CONTAINS (search_text, '$on', 1) > 0
The above query is trying stemming ($) on a stop word (on) and it is not returning any results. So, if your index has been built correctly then it should not return any rows containing the stop word.
I would try to drop and recreate the index or atleast try rebuilding it with your set of preferences again.
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
