You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by Manish Bafna <ma...@gmail.com> on 2011/09/09 15:48:52 UTC

TermsComponent from deleted document

Hi,
http://wiki.apache.org/solr/TermsComponent states that TermsComponent will
return frequencies from deleted documents too.

Is there anyway to omit the deleted documents to get the frequencies.

I know there is a facets which can be used. Is it recommended to use facets
for autosuggest feature?

Thanks,
Manish.

Re: TermsComponent from deleted document

Posted by Martijn v Groningen <ma...@gmail.com>.
I'd use the suggester:
http://wiki.apache.org/solr/Suggester

The suggester can give a collation. The TermsComponent can't do that.
The suggester builds on top of the spellchecking infrastructure, so
should be easy to use if you're familiar with that.

Martijn

On 10 September 2011 08:37, Manish Bafna <ma...@gmail.com> wrote:
>
> Which is preferable? using TermsComponent or Facets for autosuggest?
>
> On Fri, Sep 9, 2011 at 10:33 PM, Chris Hostetter
> <ho...@fucit.org>wrote:
>
> >
> > : http://wiki.apache.org/solr/TermsComponent states that TermsComponent
> > will
> > : return frequencies from deleted documents too.
> > :
> > : Is there anyway to omit the deleted documents to get the frequencies.
> >
> > not really -- until a deleted document is expunged from segment merging,
> > they are still included in the term stats which is what the TermsComponent
> > looks at.
> >
> > If having 100% accurate term counts is really important to you, then you
> > can optimize after doing any updates on your index - but there is
> > obviously a performance tradeoff there.
> >
> >
> >
> > -Hoss
> >



--
Met vriendelijke groet,

Martijn van Groningen

Re: TermsComponent from deleted document

Posted by Manish Bafna <ma...@gmail.com>.
Which is preferable? using TermsComponent or Facets for autosuggest?

On Fri, Sep 9, 2011 at 10:33 PM, Chris Hostetter
<ho...@fucit.org>wrote:

>
> : http://wiki.apache.org/solr/TermsComponent states that TermsComponent
> will
> : return frequencies from deleted documents too.
> :
> : Is there anyway to omit the deleted documents to get the frequencies.
>
> not really -- until a deleted document is expunged from segment merging,
> they are still included in the term stats which is what the TermsComponent
> looks at.
>
> If having 100% accurate term counts is really important to you, then you
> can optimize after doing any updates on your index - but there is
> obviously a performance tradeoff there.
>
>
>
> -Hoss
>

Re: TermsComponent from deleted document

Posted by Chris Hostetter <ho...@fucit.org>.
: http://wiki.apache.org/solr/TermsComponent states that TermsComponent will
: return frequencies from deleted documents too.
: 
: Is there anyway to omit the deleted documents to get the frequencies.

not really -- until a deleted document is expunged from segment merging, 
they are still included in the term stats which is what the TermsComponent 
looks at.

If having 100% accurate term counts is really important to you, then you 
can optimize after doing any updates on your index - but there is 
obviously a performance tradeoff there.



-Hoss