ft_stopword_file

[Chris Reisor]

This is with regards to MySQL 4.1.9-standard-log installed via RPM's
on Fedora Core 3.

The documentation says that changing the ft_stopword_file to a blank
variable or a custom file overrides the internal stopword list on
fulltext searches.  However, in my testing, this merely *adds to* the
stopword list, rather than *replacing* it.

How do you completely override the stopword list?  Do you just have to
compile it yourself and change myisam/ft_static.c?

Also, when trying to set the ft_stopword_file variable interactively
(from the mysql client), I get the following error:

mysql set ft_stopword_file = '';
ERROR 1193 (HY000): Unknown system variable 'ft_stopword_file'

Setting it in the my.cnf and restarting MySQL works, however.

An example of the behavior I am observing:

* Searching New England brings up all entries with england,
regardless of new.
* Changing the stoplist to blank (and doing a REPAIR TABLE)
generates the same results.
* Also changing it to a blank custom stopword list has the same results.
* Searching only on New finds no entry (even though it's in there).
* Searching on The gets the same.  Changing the custom stopword list
to include England makes the New England serach come up blank.

Thanks,
Chris

-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: ft_stopword_file

[Alec . Cawley]

 * Searching New England brings up all entries with england,
 regardless of new.
 * Changing the stoplist to blank (and doing a REPAIR TABLE)
 generates the same results.
 * Also changing it to a blank custom stopword list has the same results.
 * Searching only on New finds no entry (even though it's in there).
 * Searching on The gets the same.  Changing the custom stopword list
 to include England makes the New England serach come up blank.

You need to set the global variable ft_min_word_len to 3 to achieve what 
you want. As shipped, it is set to 4, which means that words of three or 
less letters are ignored. After changing the variable, you need to rebuild 
the index.

Alec Cawley


-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

ft_stopword_file

[Roberto Slepetys Ferreira]

Hi,

I am creating a StopWord File in Brazilian Portuguese for MySQL based in
statistical ocourrence of common words of a very large database, but it
isn't working correctly:

1) I changed the ft_stopword_file in my.conf pointing to a file where I put
the stopwords (each line one word).

2) Seeing the global variable ft_stopword_file, it points to the file that I
want.

3) Then I rebuild the indexes with REPAIR table QUICK, and dump the
dictionary: ft_dump.

4) Unhapply, the stopwords continues in the dictionary file.

Is there anything that I am missing ?

Thanks
Roberto Slepetys



-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]