Subset a GRaster layer, or return values from a column of a GVector's table
Source:R/subset_dollar.r
subset_dollar.Rd
The dollar
notation can be used to get a single layer of a multi-layer GRaster
or the values of a column from a GVector
's data table.
Arguments
- x
A
GRaster
orGVector
.- name
Character: The name of a
GRaster
or of a column of aGVector
's data table. Names of rasters and vectors can be found usingnames()
.
Examples
if (grassStarted()) {
# Setup
library(terra)
### GRasters
# Example data
madElev <- fastData("madElev") # elevation raster
madForest2000 <- fastData("madForest2000") # forest raster
madForest2014 <- fastData("madForest2014") # forest raster
# Convert SpatRasters to GRasters
elev <- fast(madElev)
forest2000 <- fast(madForest2000)
forest2014 <- fast(madForest2014)
# Re-assigning values of a GRaster
constant <- elev
constant[] <- pi
names(constant) <- "pi_raster"
constant
# Re-assigning specific values of a raster
replace <- elev
replace[replace == 1] <- -20
replace
# Subsetting specific values of a raster based on another raster
elevInForest <- elev[forest2000 == 1]
plot(c(elev, forest2000, elevInForest), nr = 1)
# Adding and replacing layers of a GRaster
rasts <- c(elev, constant, forest2000)
# Combine with another layer:
add(rasts) <- forest2014 # one way
rasts
rasts <- c(rasts, forest2014) # another way
### Subsetting GRaster layers
# Subset:
rasts <- c(elev, forest2000, forest2014)
rasts[[2:3]]
subset(rasts, 2:3)
subset(rasts, c("madForest2000", "madElev"))
rasts[[c("madForest2000", "madElev")]]
rasts$madForest2000
# Get every other layer:
rasts[[c(FALSE, TRUE)]]
### Replacing layers of a GRaster
# Replace a layer
logElev <- log(elev)
names(logElev) <- "logElev"
rasts$madForest2014 <- logElev
rasts
# Replace a layer:
rasts[[3]] <- forest2000
rasts
### GVectors
# example data
madDypsis <- fastData("madDypsis") # vector of points
# Convert SpatVector to GVector
dypsis <- fast(madDypsis)
### Retrieving GVector columns
dypsis$species # Returns the column
dypsis[[c("year", "species")]] # Returns a GRaster with these columns
dypsis[ , c("year", "species")] # Same as above
### Subsetting GVector geometries
# Subset first three geometries
dypsis[1:3]
dypsis[1:3, "species"]
# Get geometries by data table condition
dypsis[dypsis$species == "Dypsis betsimisarakae"]
### (Re)assigning GVector column values
# New column
dypsis$pi <- pi
# Re-assign values
dypsis$pi <- "pie"
# Re-assign specific values
dypsis$institutionCode[dypsis$institutionCode == "MO"] <-
"Missouri Botanical Garden"
}