[TYPO3-Solr] solr.SnowballPorterFilterFactory for multi language sites

[Ingo Renner]

Michiel Roos [netcreators] wrote:

Hi Michiel,

> Are there any known working setups for solr.SnowballPorterFilterFactory
> on multi language sites?
>
> The default is set to English. There is a Dutch lib too, which I can use
> .. . . but how about using more than one in a single config?

On http://www.be-lufthansa.com we have both, english and german content 
in one index. This works quite good as german and english are somehow 
similar up to a certain degree.
The problem of course arises with more languages and languages that are 
different from each other.
Solr currently doesn't offer a nice solution out of the box for these 
scenarios. Our idea currently is to have a multi core index setup, with 
one core per language... This way you can configure different stemmers 
for each language.
Differentiating on which index to search (and to send updates to) should 
be easy by using a TS condition that switches the path to the correct 
core in plugin.tx_solr.solr.path

For more information on Solr multicore you may want to check 
http://wiki.apache.org/solr/CoreAdmin

The best thing of course would be having Solr support different stemmers 
depending on the language of a document out of the box.


best
Ingo

-- 
Ingo Renner
TYPO3 Core Developer, Release Manager TYPO3 4.2