Re: Tsearch2 Dutch snowball stemmer in PG8.1

Alban Hertroys wrote:
> The only odd thing is that to_tsvector('dutch', 'some dutch text') now
> returns '|' for stop words...
>
> For example:
> select to_tsvector('nederlands', 'De beste stuurlui staan aan wal');
> to_tsvector
> ------------------------------------------------
> '|':1,5 'bes':2 'wal':6 'staan':4 'stuurlui':3

I found the cause. The stop words list I found contained comments
prefixed by '|' signs. Removing the contents and recreating the database
solved the problem. Just updating the reference didn't seem to help...

There's undoubtedly some cleaner way to replace the stop words list, but
at the current stage of our project this was the simplest to achieve.

--
Alban Hertroys
a(dot)hertroys(at)magproductions(dot)nl

magproductions b.v.

T: ++31(0)534346874
F: ++31(0)534346876
M:
I: www.magproductions.nl
A: Postbus 416
7500 AK Enschede

// Integrate Your World //