Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [udig-devel] Connections to R 
On Thu, 2007-03-15 at 11:27 +0100, Andrea Aime wrote:
> Adrian Custer ha scritto:
> > Hey all,
> > 
> > I'm trying to design a flexible approach to build R spatial objects.
> > Right now I'm working through JRI, trying to build a
> > SpatialPointsDataFrame in R. Neither JRI nor Rserve seem capable right
> > now of doing this out of the box, but regardless we will have to do a
> > bit of work mapping the Attribute[] array of Gt Features to a flat
> > table.
> > 
> > I gather there are different efforts working to connect udig and R,
> > either via Rserver or JRI. I believe the CGIAR potato center folk and
> > maybe axios are involved. Can any of you point me to code which builds R
> > objects? I looked on the community space and didn't find the right
> > plugins to examine.
> 
> It would be nice it this connection could a gt2 datastore.
> What about Geoserver serving data directly out of R? :-)
> Cheers
> Andrea
> _______________________________________________
> User-friendly Desktop Internet GIS (uDig)
> http://udig.refractions.net
> http://lists.refractions.net/mailman/listinfo/udig-devel

Baby steps first, eh? A datastore is the obvious end goal but it's not
clear yet how to get there. 

For now, there is too much in flux to aim for a universal datastore. You
all know about the GtFeature state but things are also evolving in R and
in the connection libararies. In particular:
        There is an experimental R effort to use a database for storage
        and sharing that database may be a better strategy for the long
        term. 

        JRI and Rserve are not mature yet so there is lots of missing
        functionality which I am trying to understand how to work around
        or implement

        R is switching over to user real classes. It's not clear yet how
        we build those classes from Java.
        
        The spatial library is only getting started. For now it uses a
        SimpleFeature like model but it's not clear how that will
        evolve.
        
so let's not get greedy from the get go ;-)

--adrian

Back to the top