[R-sig-Geo] spsample() ok with type= 'stratified', but gives error with 'random'

[Roger Bivand]

On Thu, 7 Jun 2007, Don MacQueen wrote:

> I don't know enough to assess whether I have a misunderstanding or a mistake.
> 
> spsample() with type='random' is giving an error, but succeeding with 
> type='stratified'.
> 
> Maybe random sampling requires projection information and stratified doesn't?
> I may have constructed my SpatialPolygonsDataFrame incorrectly?

The problem is that the hole well-nigh fills the polygon, and after iter=4
tries, it tries to finish, and fails, because the output object is NULL
rather than some points. Increasing iter= doesn't help. It shouldn't do
this, but it does - I'll try to fix it but not soon.

A work-around is:

plot(SP)
r1 <- spsample(as(SP, "Spatial"), type="random", n=300)
points(r1)
r2 <- overlay(SP, r1)
points(r1[!is.na(r2)], col="red")
length(which(!is.na(r2))) >= 50
r3 <- sample(which(!is.na(r2)), 50)
length(r3)
points(r1[r3], pch=19, col="red")

Most likely spsurvey and/or making an appropriate window in spatstat will 
do this better.

Thanks for a clear report that was easy to reproduce.

Roger



> 
> Thanks
> -Don
> 
> Details:
> 
> spgd consists of two rectangles, one inside the other, and the inside 
> one is a hole.
> The outer one has vertices in counter-clockwise order; the inner one 
> in clockwise order.
> 
> >  class(tst)
> [1] "SpatialPolygonsDataFrame"
> attr(,"package")
> [1] "sp"
> 
> >  summary(tst)
> Object of class SpatialPolygonsDataFrame
> Coordinates:
>          min      max
> r1 990396.8 990762.3
> r2 213369.4 213702.6
> Is projected: NA
> proj4string : [NA]
> Data attributes:
>      ID
>   zone:1
> 
> 
> >  stst <- spsample(tst,type='stratified',n=50)
> >  plot(tst)
> >  points(stst)
> >  stst <- spsample(tst,type='random',n=50)
> Error in `proj4string<-`(`*tmp*`, value = <S4 object of class "CRS">) :
>          proj4string only works for class(es extending) Spatial
> 
> 
> ##########
> >  sessionInfo()
> R version 2.5.0 (2007-04-23)
> powerpc-apple-darwin8.9.1
> 
> locale:
> C
> 
> attached base packages:
> [1] "stats"     "graphics"  "grDevices" "utils"     "datasets" 
> "methods"   "base"
> 
> other attached packages:
> maptools       sp  foreign    rmacq
> "0.6-12" "0.9-14" "0.8-20"    "1.0"
> 
> ##########
> ## since tst is fairly small, here it is:
> ##########
> >  tst
> An object of class "SpatialPolygonsDataFrame"
> Slot "data":
>         ID
> zone zone
> 
> Slot "polygons":
> [[1]]
> An object of class "Polygons"
> Slot "Polygons":
> $zn
> An object of class "Polygon"
> Slot "labpt":
> [1] 990579.4 213536.2
> 
> Slot "area":
> [1] 64601.13
> 
> Slot "hole":
> [1] FALSE
> 
> Slot "ringDir":
> [1] 1
> 
> Slot "coords":
>              x        y
> [1,] 990499.3 213702.6
> [2,] 990762.3 213559.8
> [3,] 990658.2 213369.4
> [4,] 990396.8 213513.2
> [5,] 990499.3 213702.6
> 
> 
> $zn2
> An object of class "Polygon"
> Slot "labpt":
> [1] 990584.4 213537.3
> 
> Slot "area":
> [1] 30367.20
> 
> Slot "hole":
> [1] TRUE
> 
> Slot "ringDir":
> [1] -1
> 
> Slot "coords":
>              x        y
> [1,] 990492.6 213650.8
> [2,] 990441.8 213549.4
> [3,] 990673.3 213423.7
> [4,] 990728.1 213525.6
> [5,] 990492.6 213650.8
> 
> 
> 
> Slot "plotOrder":
> [1] 1 2
> 
> Slot "labpt":
> [1] 990579.4 213536.2
> 
> Slot "ID":
> [1] "zone"
> 
> Slot "area":
> [1] 94968.33
> 
> 
> 
> Slot "plotOrder":
> [1] 1
> 
> Slot "bbox":
>          min      max
> r1 990396.8 990762.3
> r2 213369.4 213702.6
> 
> Slot "proj4string":
> CRS arguments: NA
> 
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no