Issues with Fulltext UTF search with a parser for word boundaries

[xephyris tahara]

Hi,
I have a table with UTF8 columns storing some Japanese data with a fulltext index. I have parsed the Japanese with the parser Sen and inserted spaces (" ") between each word. Some example data is:

+--------+------------------------------------------------+-----------------------+----------+
| lineno | jseg | eseg | id |
+--------+------------------------------------------------+-----------------------+----------+
| 0 | これ は テスト です | this is a test | 5 |
| 1 | テスト データ の 二 行 目 です 。 | test data second line | 5 |
| 2 | 三 行 目 です 。 | third line | 5 |
+--------+------------------------------------------------+-----------------------+----------+

Running this query:

SELECT eseg FROM segments WHERE MATCH (jseg) AGAINST ('データ')

returns:
Empty set (0.00 sec)

Is this the expected behaviour?

I'm inserting a space character as the word boundary, maybe I should use something else?
I realise Senna is available, but would like to know why this isn't working first :)



Edited 1 time(s). Last edit at 04/18/2008 10:49PM by xephyris tahara.