Re: Is SOLR best suited to this application - Finding co-ordinates

[Lance Norskog <go...@gmail.com>]

How many unique places, coordinates and results do you have? Solr
works very well when you denormalize your data into one document per
searchable combination. It can handle tens of millions of these
records without a problem.

On Tue, Jul 31, 2012 at 3:17 AM, Spadez <ja...@hotmail.com> wrote:
> I already have a SOLR server up and running which is designed to receive
> keywords and co-ordinates and result item results.
>
> The way I see it I have three options:
>
> 1. Have one "geocode" solr server to convert place names to co-ordinates and
> another seperate Solr server to turn the items results.
> 2. Use the same SOLR server for both - one index for items and co-ordinates,
> another index for placenames and co-ordinates.
> 3. Add some sort of plugin / addon to perform the geocode conversion inline
> when the query is received.
>
> Which seems logical?
>
>
>
> --
> View this message in context: http://lucene.472066.n3.nabble.com/Is-SOLR-best-suited-to-this-application-Finding-co-ordinates-tp3998308p3998323.html
> Sent from the Solr - User mailing list archive at Nabble.com.



-- 
Lance Norskog
goksron@gmail.com

Re: Is SOLR best suited to this application - Finding co-ordinates

[Spadez <ja...@hotmail.com>]

Normalising the data is a good idea, and it would be easy to do since I would
only have around 50,000 entires BUT it is a bit complicated with addresses I
think. Lets say I store the data in this form:

<Town><City><Country>

London, England
Swindon, Wiltshire, England
Wiltshire England
England

What happens if someone searches just "London", or just "Swindon". I assume
it wouldnt return any results because they would have to type London,
England for example. If I include an entry for London and "London, England"
then the autocomplete will show both, which would confuse the user.




--
View this message in context: http://lucene.472066.n3.nabble.com/Is-SOLR-best-suited-to-this-application-Finding-co-ordinates-tp3998308p3998547.html
Sent from the Solr - User mailing list archive at Nabble.com.