Mailing List Archive

Just save the searches in a table on a DB.

-----Original Message-----
From: Ryan Hutchins [mailto:ryan@myvidster.com]
Sent: Thursday, August 06, 2015 9:32 AM
To: general@lucene.apache.org
Subject: solr search

I'm looking for a way to find out my most searched terms using solr. Is this possible. I've looked on line and haven't seen much in the way of solutions.

*-Ryan*

*Webmaster*
myVidster.com
ICQ: 679709664

Thank you for your reply. I thought of this. My site has about 600,000
active users. I'm no SQL expert but wouldn't that put an incredible strain
on the server?


*-Ryan*

*Webmaster*
myVidster.com
ICQ: 679709664

On Thu, Aug 6, 2015 at 11:35 AM, Doug Ford <DFord@nu.edu> wrote:

> Just save the searches in a table on a DB.
>
> -----Original Message-----
> From: Ryan Hutchins [mailto:ryan@myvidster.com]
> Sent: Thursday, August 06, 2015 9:32 AM
> To: general@lucene.apache.org
> Subject: solr search
>
> I'm looking for a way to find out my most searched terms using solr. Is
> this possible. I've looked on line and haven't seen much in the way of
> solutions.
>
> *-Ryan*
>
> *Webmaster*
> myVidster.com
> ICQ: 679709664
>

> Thank you for your reply. I thought of this. My site has about 600,000
> active users. I'm no SQL expert but wouldn't that put an incredible strain
> on the server?

Hi Ryan,

The real question is how many searches per minute are happening, and
whether adding that many INSERTs will be a problem. If it is, you could
instead increment the counts in a faster-to-update store like Redis, and
then every 10 minutes (or whatever) flush that to your database. Note
none of this is really depends on Solr or Lucene. :-)

Good luck!

Paul

On Thu, Aug 6, 2015, at 05:48 PM, Paul Jungwirth wrote:
> > Thank you for your reply. I thought of this. My site has about 600,000
> > active users. I'm no SQL expert but wouldn't that put an incredible strain
> > on the server?
>
> Hi Ryan,
>
> The real question is how many searches per minute are happening, and
> whether adding that many INSERTs will be a problem. If it is, you could
> instead increment the counts in a faster-to-update store like Redis, and
> then every 10 minutes (or whatever) flush that to your database. Note
> none of this is really depends on Solr or Lucene. :-)

Or if the number of searches done is smaller, you have a perfectly good
datastore in Solr/Lucene. You can find the most used searches using
faceting and other familiar methods.

Upayavira

Log all the searched query and process the log offline to find number of
searches and frequency.

On 6 August 2015 at 22:27, Upayavira <uv@odoko.co.uk> wrote:

>
>
> On Thu, Aug 6, 2015, at 05:48 PM, Paul Jungwirth wrote:
> > > Thank you for your reply. I thought of this. My site has about 600,000
> > > active users. I'm no SQL expert but wouldn't that put an incredible
> strain
> > > on the server?
> >
> > Hi Ryan,
> >
> > The real question is how many searches per minute are happening, and
> > whether adding that many INSERTs will be a problem. If it is, you could
> > instead increment the counts in a faster-to-update store like Redis, and
> > then every 10 minutes (or whatever) flush that to your database. Note
> > none of this is really depends on Solr or Lucene. :-)
>
> Or if the number of searches done is smaller, you have a perfectly good
> datastore in Solr/Lucene. You can find the most used searches using
> faceting and other familiar methods.
>
> Upayavira
>