I have been having some strange error messages from the rasterFromXYZ function in the R raster package. Here is an example
library(raster)
xyz <- data.frame(x = c(5.463636, 5.481818, 5.5), y = c(51.42727, 51.42727, 51.42727), z = c(1.2,1.3,1.6))
r <- rasterFromXYZ(xyz)
##error
Error in if (nc > (2^31 - 1)) return(FALSE) :
missing value where TRUE/FALSE needed
In addition: Warning message:
In min(dy) : no non-missing arguments to min; returning Inf
##specifying the resolution as 1
r <- rasterFromXYZ(xyz, res = 1)
##different error
Error in rasterFromXYZ(xyz, res = 1) : x cell sizes are not regular
The x coordinates are perfectly regular. What am I doing wrong?
The x-coordinates are OK, but there is only one unique y-coordinate value. So there is no way to guess the vertical resolution.
xyz
# [,1] [,2] [,3]
#[1,] 5.463636 51.42727 1.2
#[2,] 5.481818 51.42727 1.3
#[3,] 5.500000 51.42727 1.6
If you set the resultion to 1 that does not match the x-coordinates, but you can do
rasterFromXYZ(xyz, res=c(NA, 1))
#class : RasterLayer
#dimensions : 1, 3, 3 (nrow, ncol, ncell)
#resolution : 0.018182, 1 (x, y)
#extent : 5.454545, 5.509091, 50.92727, 51.92727 (xmin, xmax, ymin, ymax)
#crs : NA
#source : memory
#names : layer
#values : 1.2, 1.6 (min, max)
The development version now gives a better error message:
r <- rasterFromXYZ(xyz)
#Error in rasterFromXYZ(xyz) : more than one unique y value needed