Joining multicore to return top results

[Selvam <s....@gmail.com>]

Hi,

This should be trivial question, still I am failing to get the details.I
have 2 cores+default collection,

*collection1:*
article_id
title
content

*core0:*
cluster_id
cluster_name
cluster_count

*core1:*
article_id
article_cluster_id
score

Given an article_id, I want to return top 10 ( based on score field in
core1 ) other articles falling in the same cluster. I would like to know,
how to implement this as I am fairly new to Solr.

Version used : Solr 3.5
-- 
Regards,
S.Selvam
http://knackforge.com

Re: Joining multicore to return top results

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

>From  your example, it rather looks like you've moved some DB
tables into separate cores and are trying to do some SQL-like
operations. Stop that! <G>...

Solr really isn't built for this kind of operation. I know this goes
against all your DB training, but can you simply de-normalize
all the data and put it into a single core/index? Solr searches lots and
lots of data very efficiently, but any time you try to make it behave
like a database, you should take a step back and ask if this is
really the way to use Solr...

And, of course, I may be misinterpreting your problem entirely...

Best
Erick

On Fri, Feb 10, 2012 at 11:35 PM, Selvam <s....@gmail.com> wrote:
> Hi,
>
> This should be trivial question, still I am failing to get the details.I
> have 2 cores+default collection,
>
> *collection1:*
> article_id
> title
> content
>
> *core0:*
> cluster_id
> cluster_name
> cluster_count
>
> *core1:*
> article_id
> article_cluster_id
> score
>
> Given an article_id, I want to return top 10 ( based on score field in
> core1 ) other articles falling in the same cluster. I would like to know,
> how to implement this as I am fairly new to Solr.
>
> Version used : Solr 3.5
> --
> Regards,
> S.Selvam
> http://knackforge.com