How to select which layers I want to keep in a new SpatRaster

rrasterterra

I have a raster with 10 layers. I want to split it up into two separate objects, one containing the first 5 layers, and one containing the second 5 layers. Reading the help file for rast(), it seems like when I create a new SpatRaster I can select the NUMBER of layers that will be retained from the original, but I don't see how I can specify WHICH layers will be retained from the original. If I just specify 5 layers, is it only keeping the first 5 layers?

x <- rast(nlyrs = 10, nrows=108, ncols=21, xmin=0, xmax=10)

y <- rast(x, nlyrs = 5)

Best Answer

You can simply index the rasters. In this example we will subset the (1, 3-4, 10) rasters. Note the double brackets.

library(terra)
x <- rast(nlyrs = 10, nrows=108, ncols=21, xmin=0, xmax=10)
  y <- x[[c(1, 3:4, 10)]]

For your needs you can use

y1 <- x[[1:5]]
y2 <- x[[6:10]]

Unlike the raster package, to add a raster you can simply use concatenate.

y <- c(y, x[[2]])

There are also some formal methods in terra, see; subset, selectRange and, add

Related Solutions

[GIS] Joining attributes by location for more than one located feature using R

You can use raster::union

Example data:

library(tmap)
library(rgeos)
library(rgdal)
library(raster)
data(Europe) 
# simplifying a bit
Eu <- Europe[Europe$iso_a3 %in% c('ESP', 'PRT', 'FRA'),1:2]
data(metro)
metro.pts <- spTransform(metro[, 'name'], crs(Europe))
metro.poly <- gBuffer(metro.pts, width=50000, byid=TRUE)

Original solution

u <- union(Eu, metro.poly)

i <- intersect(Eu, metro.poly)

**Update, now I understand the question better **

You can returnList = TRUE in over:

v <- over(Eu, metro.poly, returnList=TRUE)

You could now do:

a <- lapply(1:length(v), function(i) cbind(iso_a3=Eu$iso_a3[i], v[[i]]))
b <- do.call(rbind, a)
m <- merge(Eu, b, by="iso_a3", all.x=TRUE, duplicateGeoms=TRUE)

In merge, the argument duplicateGeoms=TRUE is necessary because otherwise merge cannot handle having more than one (i.e. duplicate) value to assign to each feature.
To get the following:

data.frame(m)
#  iso_a3   name.x    name.y
#1    ESP    Spain Barcelona
#2    ESP    Spain    Madrid
#3    FRA   France     Lille
#4    FRA   France      Lyon
#5    FRA   France Marseille
#6    FRA   France     Paris
#7    PRT Portugal    Lisbon
#8    PRT Portugal     Porto

I think that this is the same as what Tiernan showed in perhaps more elegant sf code, but I do not think that this is what you asked for.

For what you want, we need to create a data.frame from a list with varying number of elements. There might be a better way, but here is one:

m <- max(sapply(v, nrow))
iso <- as.character(Eu$iso_a3)

x <- sapply(1:length(v), 
       function(i) c(iso[i], c(v[[i]][,1], rep(NA, m))[1:m])
     )

x <- data.frame(t(x), stringsAsFactors=FALSE)
colnames(x) <- c("iso_a3", paste0("metro_", 1:m))

Merge with the original polygons:

r <- merge(Eu, x, by="iso_a3", all.x=TRUE)
data.frame(r)

#  iso_a3     name   metro_1 metro_2   metro_3 metro_4
#1    ESP    Spain Barcelona  Madrid      <NA>    <NA>
#2    FRA   France     Lille    Lyon Marseille   Paris
#3    PRT Portugal    Lisbon   Porto      <NA>    <NA>

Raster and polygon layer at different location

You state that you deleted the header of the file. That's why it doesn't work because the header includes the CRS definition: Koordinatensystem=Gauss-Krueger 3. Meridianstreifen Potsdam-Datum. This corresponds to EPSG:31467 (as is stated in the documentation to the data, to be found on the site you linked to).

What you have to do: right-click your layer / Layer CRS / Set layer CRS and select EPSG:31467.

Best Answer

Related Solutions

[GIS] Joining attributes by location for more than one located feature using R

Raster and polygon layer at different location

Related Question