Convert a SpatialPolygonsDataFrame to raster using rasterize function

Question

I'm trying to convert my map (a SpatialPolygonsDataFrame) into a raster, using rasterize function, here is a bug that I don't understand...

> class(parcel) # this is my map
[1] "SpatialPolygonsDataFrame"

> attr(,"package")
[1] "sp"

> head(parcel)
AREAHA RECNO      AREA AREASQKM        OS_06 GRD_CAT_06
0 0.0519     0  519.1941        0 chemin prive        way
1 0.0127     1  127.0963        0 chemin prive        way
2 0.0596     2  596.2204        0 chemin prive        way
3 0.0248     3  248.0452        0    route D90        way
4 0.0375     4  375.0141        0 route privee        way
5 0.1260     5 1260.3941        0 chemin prive        way

> dim(parcel)
[1] 20069     6

> object.size(parcel)
88090144 bytes

> r <- raster(ncol=4001, nrow=4001)
> extent(r) <- extent(parcel)
> parcel_ras<-rasterize(parcel, r, "GRD_CAT_06", CRS("+init=epsg:27572"))
Erreur dans data.frame(ID = as.integer(putvals), value = lvs) :    arguments imply differing number of rows: 20069, 13

I just don't understand this comment : what is "putvals" ? There are 20069 polygons in my map, so what is "13" ?

I can't reproduce a smaller example of my map, so here are my data : http://dl.free.fr/fxdT2oczB

Answer 1

Your error says as.integer(putvals). The R rasterize function can't work on Strings. You have to transform your data first.

Something like this may work, but i would assign different ranks (ala 1,2,3,4,..) to your data. However i still get a (different) error for which i don't have an explanation. Maybe the size of your raster is incorrect...

library(sp);library(rgdal);library(raster)

r <- raster(ncol=4001, nrow=4001)

extent(r) <- extent(parcel)

parcel$Grd_ranks <- rank(parcel$GRD_CAT_06)

parcel_ras<- rasterize(parcel, r, "Grd_ranks", CRS("+init=epsg:27572"),fun='first')

Maybe you should use the QGIS or GRASS rasterize function instead. Much less complicated ;-)