[R-sig-Geo] New (?) use of polypath() in package sp causing a problem?

[Roger Bivand]

On Fri, 10 Aug 2012, MacQueen, Don wrote:

> I have encountered a problem I've never seen before.
>
> The problem is:
>
>> x11(type='Xlib')
>> plot(sfbound)
> Warning message:
> In polypath(x = mcrds[, 1], y = mcrds[, 2], border = border, col = col, :
> Path drawing not available for this device
>
> And nothing is plotted.
>
> The object sfbound is a SpatialPolygonsDataFrame object with one simple
> polygon in it.
> Previously I have had no problem plotting such objects with
> x11(type='Xlib').x11(type='cairo') succeeds.
>
>
> It looks very likely that this relates to this entry in NEWS for R 2.12.0:
>
> -- quote --
> Added support for polygons with holes to the graphics
> engine. This is implemented for the pdf(),
> postscript(), x11(type="cairo"), windows(),
> and quartz() devices (and associated raster formats),
> but not for x11(type="Xlib") or xfig() or
> pictex(). The user-level interface is the
> polypath() function in graphics and
> grid.path() in grid.
> -- end quote --
>
> That is, presuming a relatively recent change in sp in which polyplot() is
> now used and previously was not, the change described in NEWS would
> explain why I'm seeing this.
>
>
>
> The problem is, I have found that for plots in which a very large number
> of objects are being plotted, the Xlib version of X11() is much faster
> than the cairo version. And I frequently plot spatial objects with
> thousands of line segments in them, where the performance difference is
> substantial, to the point of making use of the cairo version intolerable.
>
> I guess the question is -- am I correct in this analysis?
> If so, do I have any alternative when I need to plot spatial objects with
> huge numbers of points/lines/polygons in them?

Don,

The analysis is correct (r1248). If the hole status of Polygon objects is 
known to be correct, and the par(bg=) and pbg= values are set to match, 
the plots will be the same. polypath analyses the polygon rings itself and 
provides transparency in polygon painting - it doesn't paint in holes. The 
legacy polygon function first painted the whole exterior ring, then 
overpainted holes with the pbg= colour, which is "transparent" by default. 
For many purposes, polypath is preferable, but as you know, cairo is much 
slower than X11. So using X11 ought still to be OK, in the plot method set 
usePolypath=FALSE. We can add back an option and set the argument by 
default to that option if that would help.

The details for the plot method in the code:

https://r-forge.r-project.org/scm/viewvc.php/pkg/sp/R/SpatialPolygons-displayMethods.R?root=rspatial&view=log

and for spplot methods look in the same repository.

Hope this helps,

Roger

>
> Thanks
> -Don
>
>
>> str(sfbound)
> Formal class 'SpatialPolygonsDataFrame' [package "sp"] with 5 slots
>  ..@ data       :'data.frame': 1 obs. of  1 variable:
>  .. ..$ name: Factor w/ 1 level "Boundary.kml": 1
>  ..@ polygons   :List of 1
>  .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
>  .. .. .. ..@ Polygons :List of 1
>  .. .. .. .. ..$ :Formal class 'Polygon' [package "sp"] with 5 slots
>  .. .. .. .. .. .. ..@ labpt  : num [1:2] -121.7 37.7
>  .. .. .. .. .. .. ..@ area   : num 3.71e-06
>  .. .. .. .. .. .. ..@ hole   : logi FALSE
>  .. .. .. .. .. .. ..@ ringDir: int 1
>  .. .. .. .. .. .. ..@ coords : num [1:8, 1:2] -122 -122 -122 -122 -122
> ...
>  .. .. .. ..@ plotOrder: int 1
>  .. .. .. ..@ labpt    : num [1:2] -121.7 37.7
>  .. .. .. ..@ ID       : chr "1"
>  .. .. .. ..@ area     : num 3.71e-06
>  ..@ plotOrder  : int 1
>  ..@ bbox       : num [1:2, 1:2] -121.7 37.7 -121.7 37.7
>  .. ..- attr(*, "dimnames")=List of 2
>  .. .. ..$ : chr [1:2] "x" "y"
>  .. .. ..$ : chr [1:2] "min" "max"
>  ..@ proj4string:Formal class 'CRS' [package "sp"] with 1 slots
>  .. .. ..@ projargs: chr "+proj=longlat +ellps=GRS80 +datum=NAD83
> +no_defs +towgs84=0,0,0"
>
>
>> sessionInfo()
> R version 2.15.1 (2012-06-22)
> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
>
> locale:
> [1] C
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> other attached packages:
> [1] sp_0.9-99   rmacq_1.1-8
>
> loaded via a namespace (and not attached):
> [1] grid_2.15.1    lattice_0.20-6 tools_2.15.1
>
>
>
>

-- 
Roger Bivand
Department of Economics, NHH Norwegian School of Economics,
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