PorterStemmer using example schema and data

["Birkmann, Magdalena" <ma...@open-xchange.com>]

Hey there,
I've been working through the Solr Tutorial
(http://lucene.apache.org/solr/tutorial.html), using the example schema and
documents, just working through step by step trying everything out. Everything
worked out the way it should (just using the example queries and stuff), except
for the stemming (A search for features:recharging
<http://localhost:8983/solr/select/?indent=on&q=features:recharging&fl=name,features>
should match Rechargeable due to stemming with the EnglishPorterFilter, but
doesn't). I've been the using the example directory exactly the way it was when
downloading it, without changing anything. Since I'm fairly new to all of this
and don't quite understand yet how all of it works or should work, I don't
really know where the problem lies or how to configure anything to make it work,
so I just thought I'd ask here, since you all seem so nice :)
Thanks a lot in advance,
Magda

Re: PorterStemmer using example schema and data

["Birkmann, Magdalena" <ma...@open-xchange.com>]

I tried that, and it seems like "recharging" and "rechargeable", for example,
actually do stem to the same root (recharg). So why is it not working when I'm
searching on my indexed sampledocs? The stemming works when I search for
"videos" and it's actually "video" in the document, etc., but not for
rechargeable-recharging or capability-capable, etc., even though they stem to
the same root when i check them on the Admin/analysis page. What am I
overlooking?


On March 16, 2012 at 2:17 PM Erick Erickson <er...@gmail.com> wrote:

> What you think the results of stemming should be and what they
> actually are sometimes differ <G>...
>
> Look at the admin/analysis page, check the "verbose" boxes
> and try recharging rechargeable and you'll see, step by step,
> the results of each element of the analysis chain. Since
> the Porter stemmer is algorithmic, I'm betting that
> these don't stem to the same root.
>
> Best
> Erick
>
> On Thu, Mar 15, 2012 at 7:05 AM, Birkmann, Magdalena
> <ma...@open-xchange.com> wrote:
> >
> > Hey there,
> > I've been working through the Solr Tutorial
> > (http://lucene.apache.org/solr/tutorial.html), using the example schema and
> > documents, just working through step by step trying everything out.
> > Everything
> > worked out the way it should (just using the example queries and stuff),
> > except
> > for the stemming (A search for features:recharging
> > <http://localhost:8983/solr/select/?indent=on&q=features:recharging&fl=name,features>
> > should match Rechargeable due to stemming with the EnglishPorterFilter, but
> > doesn't). I've been the using the example directory exactly the way it was
> > when
> > downloading it, without changing anything. Since I'm fairly new to all of
> > this
> > and don't quite understand yet how all of it works or should work, I don't
> > really know where the problem lies or how to configure anything to make it
> > work,
> > so I just thought I'd ask here, since you all seem so nice :)
> > Thanks a lot in advance,
> > Magda

Re: PorterStemmer using example schema and data

[Erick Erickson <er...@gmail.com>]

What you think the results of stemming should be and what they
actually are sometimes differ <G>...

Look at the admin/analysis page, check the "verbose" boxes
and try recharging rechargeable and you'll see, step by step,
the results of each element of the analysis chain. Since
the Porter stemmer is algorithmic, I'm betting that
these don't stem to the same root.

Best
Erick

On Thu, Mar 15, 2012 at 7:05 AM, Birkmann, Magdalena
<ma...@open-xchange.com> wrote:
>
> Hey there,
> I've been working through the Solr Tutorial
> (http://lucene.apache.org/solr/tutorial.html), using the example schema and
> documents, just working through step by step trying everything out. Everything
> worked out the way it should (just using the example queries and stuff), except
> for the stemming (A search for features:recharging
> <http://localhost:8983/solr/select/?indent=on&q=features:recharging&fl=name,features>
> should match Rechargeable due to stemming with the EnglishPorterFilter, but
> doesn't). I've been the using the example directory exactly the way it was when
> downloading it, without changing anything. Since I'm fairly new to all of this
> and don't quite understand yet how all of it works or should work, I don't
> really know where the problem lies or how to configure anything to make it work,
> so I just thought I'd ask here, since you all seem so nice :)
> Thanks a lot in advance,
> Magda