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 Varun Gupta <va...@gmail.com> on 2009/09/29 11:36:27 UTC

Showing few results for each category (facet)

Hi,

I am looking for a way to do the following in solr:
When somebody does a search, I want to show results by category (facet) such
that I display 5 results from each category (along with showing the total
number of results in each category which I can always do using the facet
search). This is kind of an overview of all the search results and user can
click on the category to see all the results pertaining to that category
(normal facet search with filter).

One way that I can think of doing this is by making as many queries as there
are categories and show these results under each category. But this will be
very inefficient. Is there any way I can do this ?

Thanks & Regards,
Varun Gupta

Re: Showing few results for each category (facet)

Posted by Varun Gupta <va...@gmail.com>.

Thanks Matt!! I will take a look at the patch for field collapsing.

Thanks Marian for pointing that out. If the field collapse does not work
then I will have to rely on solr caching.

Thanks,
Varun Gupta


On Wed, Sep 30, 2009 at 1:44 AM, Matt Weber <ma...@mattweber.org> wrote:

> So, you want to display 5 results from each category and still know how
> many results are in each category.  This is a perfect situation for the
> field collapsing patch:
>
> https://issues.apache.org/jira/browse/SOLR-236
> http://wiki.apache.org/solr/FieldCollapsing
>
> Here is how I would do it.
>
> Add a field to your schema called category or whatever.  Then while
> indexing you populate that field with whatever category the document belongs
> in.  While executing a search, collapse the results on that field with a max
> collapse of 5.  This will give you at most 5 results per category.  Now, at
> the same time enable faceting on that field and DO NOT use the collapsing
> parameter to recount the facet vales.  This means that the facet counts will
> be reflect the non-collapsed results.  This facet should only be used to get
> the count for each category, not displayed to the user.  On your search
> results page that gets the collapsed results, you can put a link that says
> "Show all X results from this category" where X is the value you pull out of
> the facet.  When a user clicks that link you basically do the same search
> with field collapsing disabled, and a filter query on the specific category
> they want to see, for example:  &fq=category:people.
>
> Hope this helps.
>
> Thanks,
>
> Matt Weber
>
>
> On Sep 29, 2009, at 4:55 AM, Marian Steinbach wrote:
>
>  On Tue, Sep 29, 2009 at 11:36 AM, Varun Gupta <va...@gmail.com>
>> wrote:
>>
>>> ...
>>>
>>> One way that I can think of doing this is by making as many queries as
>>> there
>>> are categories and show these results under each category. But this will
>>> be
>>> very inefficient. Is there any way I can do this ?
>>>
>>
>>
>> Hi Varun!
>>
>> I think that doing multiple queries doesn't have to be inefficient,
>> since Solr caches subsequent queries for the same term and facets.
>>
>> Imagine this as your first query:
>> - q: xyz
>> - facets: myfacet
>>
>> and this as a second query:
>> - q:xyz
>> - fq: myfacet=a
>>
>> Compared to the first query, the second query will be very fast, since
>> all the hard work ahs been done in query one and then cached.
>>
>> At least that's my understanding. Please correct me if I'm wrong.
>>
>> Marian
>>
>
>

Re: Showing few results for each category (facet)

Posted by Matt Weber <ma...@mattweber.org>.

So, you want to display 5 results from each category and still know  
how many results are in each category.  This is a perfect situation  
for the field collapsing patch:

https://issues.apache.org/jira/browse/SOLR-236
http://wiki.apache.org/solr/FieldCollapsing

Here is how I would do it.

Add a field to your schema called category or whatever.  Then while  
indexing you populate that field with whatever category the document  
belongs in.  While executing a search, collapse the results on that  
field with a max collapse of 5.  This will give you at most 5 results  
per category.  Now, at the same time enable faceting on that field and  
DO NOT use the collapsing parameter to recount the facet vales.  This  
means that the facet counts will be reflect the non-collapsed  
results.  This facet should only be used to get the count for each  
category, not displayed to the user.  On your search results page that  
gets the collapsed results, you can put a link that says "Show all X  
results from this category" where X is the value you pull out of the  
facet.  When a user clicks that link you basically do the same search  
with field collapsing disabled, and a filter query on the specific  
category they want to see, for example:  &fq=category:people.

Hope this helps.

Thanks,

Matt Weber

On Sep 29, 2009, at 4:55 AM, Marian Steinbach wrote:

> On Tue, Sep 29, 2009 at 11:36 AM, Varun Gupta  
> <va...@gmail.com> wrote:
>> ...
>>
>> One way that I can think of doing this is by making as many queries  
>> as there
>> are categories and show these results under each category. But this  
>> will be
>> very inefficient. Is there any way I can do this ?
>
>
> Hi Varun!
>
> I think that doing multiple queries doesn't have to be inefficient,
> since Solr caches subsequent queries for the same term and facets.
>
> Imagine this as your first query:
> - q: xyz
> - facets: myfacet
>
> and this as a second query:
> - q:xyz
> - fq: myfacet=a
>
> Compared to the first query, the second query will be very fast, since
> all the hard work ahs been done in query one and then cached.
>
> At least that's my understanding. Please correct me if I'm wrong.
>
> Marian

Re: Showing few results for each category (facet)

Posted by Marian Steinbach <ma...@sendung.de>.

On Tue, Sep 29, 2009 at 11:36 AM, Varun Gupta <va...@gmail.com> wrote:
> ...
>
> One way that I can think of doing this is by making as many queries as there
> are categories and show these results under each category. But this will be
> very inefficient. Is there any way I can do this ?

Hi Varun!

I think that doing multiple queries doesn't have to be inefficient,
since Solr caches subsequent queries for the same term and facets.

Imagine this as your first query:
- q: xyz
- facets: myfacet

and this as a second query:
- q:xyz
- fq: myfacet=a

Compared to the first query, the second query will be very fast, since
all the hard work ahs been done in query one and then cached.

At least that's my understanding. Please correct me if I'm wrong.

Marian