Hi,
I have recompiled mysql using the hungarian character set, reindexed all
the tables,...
I have strange problem now:
If I do:
select count(*)
from sites
where match(site_content)
against('tőzsde');
it returns 0.
If I do:
select count(*)
from sites
where site_content like "% tőzsde %";
it returns 38.
Of course tőzsde cannot be a stopword, as it is long enough, and this
table contains some 10000 records, and only a very few of them is
relevant.
The other interesting thing is, that it works for example with the
string "értéktőzsde", match-againt finds the relevant records.
See the previous conditions below.
Regards,
Peter
Peter Szekszardi wrote:
>
> Hi,
>
> I have configured MySQL with charset latin2 on a Linux box. I made a big
> table with a mediumtext filed and a search engine using MySQL
> match-against. On the Linux box everything looked fine. Now we replaced
> our server to FreeBSD. I compiled the same source with the same
> configuration settings. Now I have several problems with special
> iso-latin2 characters:
> - For example when I search for a string which contains a character
> which has the code of F5 (t+F5+zsde), it selects words which is the same
> but contains the character with code F4 (t+F4+zsde). The difference is
> that the character F4 is an element of the standard character set also.
> They look familiar and being afraid of codepage problems some sites use
> F4 instead of the official F5. A nice behaviour from MySQL that it also
> select the words with F6 (t+F6+zsde). F5 and F6 completely different
> letters, although they pronounced nearly the same and the look of letter
> itself is similar. Again, F6 is the member of the standard codepage.
> - Another example if I search for a word starts with F5 (F5+rmester), I
> get no results from MySQL. If I use F6 (F6+rmester) which is
> grammaticaly completely incorrect, I get the result. This is funny
> because the result contains F5 (F5+rmester)!
> (F5 is o with two long diacritics, F4 is o with one small v turned
> upsidedown on the top of it, F6 is o with two dots on it. F5 & F6 are
> the official characters, but they are different.)
>
> I would really apreciate if someone could enlighten me... Where is the
> problem? Is there a problem wih MySQL? Or the problem is with the
> localisation function of FreeBSD? Both the previous Linux server and the
> current BSD has setlocale installed and has the iso-8859.2 locale.
> I am sorry, but my English might be to week to explain the situation
> correctly.
>
> Regards,
>
> Peter
---------------------------------------------------------------------
Before posting, please check:
http://www.mysql.com/manual.php (the manual)
http://lists.mysql.com/ (the list archive)
To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
