Sorry, my bad on my 2nd point - Solr does
support arbitrary n-dimensional space using PointType.

So that leaves my only problem with "not recommended for production use at
this time." I think I will give it a shot, and hopefully it won't explode
on me too badly!

Thank you for the recommendations.
Ryan



On Sun, May 19, 2013 at 10:02 PM, Ryan Chazen <[email protected]> wrote:

> Hi Eric
>
> The warning on Yokozuna looks a bit troublesome... " It is in
> developmental stages and is not recommended for production use at this
> time."
>
> Also, Solr requires lat/long data, and my data is X,Y coordinates, so the
> lat/long calculations/boundaries would possibly be incorrect... although it
> might be possible to work around this with some kind of transformation. The
> solution itself seems a bit unwieldy and difficult to maintain though, so I
> don't think I'd want to use it as a first choice...
>
> Ryan
>
>
>
> On Sun, May 19, 2013 at 9:03 PM, Eric Redmond <[email protected]> wrote:
>
>> Rather than geohashing, you might consider Yokozuna. It has built in geo
>> searches.
>>
>> Eric
>>
>> On May 19, 2013, at 11:47 AM, Ryan Chazen <[email protected]> wrote:
>>
>> > Hi
>> >
>> > I'm looking at Riak as a solution for handling this type of data:
>> >
>> > A map is divided into sectors, and on the map there are many objects
>> with each object being in 1 or more sectors. (An object can be in more than
>> 1 sector if it is larger than the sector). Objects can move, so object 323
>> could move from sector 5 to sector 10.
>> >
>> > Queries come in of the form : Retrieve all objects in a list of
>> sectors. eg Sectors[55, 378, 272, ...] - maximum ~50 sectors per request.
>> >
>> > From looking through Riak docs, 2i indexes would be perfect but they
>> only seem to allow range queries, and not a list of ids.
>> > Multiple range queries could be used, but I'm guessing this will be
>> heavy on the servers?
>> >
>> > Map-Reduce looked like a good fit, but the warning on the page states
>> "When you want to query data of an entire bucket. MapReduce uses a list of
>> keys, which can place a lot of demand on the cluster." - so I guess
>> MapReduce is out?
>> >
>> > The fulltext search seems possible, but I'm a bit confused if it will
>> work for this.
>> >
>> >
>> > Any recommendations, or is Riak not a good fit for this type of data?
>> >
>> > Regards,
>> > Ryan
>> >
>> > _______________________________________________
>> > riak-users mailing list
>> > [email protected]
>> > http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>>
>
>
_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com

Reply via email to