[R-sig-Geo] R 3.0.0 and spatial classes

[Barry Rowlingson]

On Sun, Dec 16, 2012 at 6:53 PM, Chris English <sglish at hotmail.com> wrote:
>
> Edzer:
> On Tue, Jul 17, 2012 at 2:34 PM, Agustin Lobo ~ from Barry Rowlingson
> wrote:
> To convert to SpatialLines, get the coordinates and build in the
> usual convoluted manner:
>
>  > s=data.frame(x=runif(10),y=1:10,z=rnorm(10))
>  > coordinates(s)=~x+y
>  > L = SpatialLines(list(Lines(list(Line(coordinates(s))),"X")))
>  > plot(L)
> Convoluted is not the same as orphaned, certainly, but one gets the sense that
> 'Line' owes its existence to matters of plotting rather than line as line, independent of
> drawing it, and this may have some import upon line analysis and the possibility of
> arriving at topology and dispensing with shared lines and the like.

 "Line" is a non-spatial Line, made up from an ordered set of (x,y)
coordinate pairs. It cannot have a coordinate system assigned to it.

 "Lines" is a list of Lines, making a non-spatial set of Line
segments. It too cannot have a coordinate system assigned.

 "SpatialLines" is a spatial set of Lines, for when you have a number
of features each of which may be composed of several disconnected
segments. It can have a coordinate system assigned.

 You are making a SpatialLines object where a single feature has a
single line segment. The slightly annoying
list(Lines(list(Line(...)))) dance is unavoidable because the lists
are necessary since the lists can have more than one element, but you
could easy create a 'SimpleSpatialLine' function that did all that. It
might even make sp for R 3.0.0! (SimpleSpatialPolygon might be handy
too...)

Another possibility might be to write methods for SpatialLines that
takes a matrix and skips the complexity for simple cases...

Barry