Extract all values as one column of a raster?-CodePudding

If I have this stack with three layers.

     library(terra)  
     y <- rast(ncol=10, nrow=10, nlyr=3, vals=rep(1:3, each=100))

I would like to extract the values in one column as follows:

 pixels 1 from lyr.1
 pixels 1 from lyr.2
 pixels 1 from lyr.3
 pixels 2 from lyr.1
 pixels 2 from lyr.2
 pixels 2 from lyr.3

etc

CodePudding user response：

You can convert the SpatRaster to a data frame with the cell index and then manipulate it (e.g. with tidyverse) to suit your needs:

library(terra)  
#> terra 1.5.34
y <- rast(ncol=10, nrow=10, nlyr=3, vals=rep(1:3, each=100))

df <- as.data.frame(y, cells=TRUE)

# Tidyverse approach
library(tidyverse)
final_df <- df %>% pivot_longer(cols=c(lyr.1, lyr.2, lyr.3),
                    names_to = "layer") %>%
  rename(pixel = cell)

final_df
#> # A tibble: 300 x 3
#>    pixel layer value
#>    <int> <chr> <int>
#>  1     1 lyr.1     1
#>  2     1 lyr.2     2
#>  3     1 lyr.3     3
#>  4     2 lyr.1     1
#>  5     2 lyr.2     2
#>  6     2 lyr.3     3
#>  7     3 lyr.1     1
#>  8     3 lyr.2     2
#>  9     3 lyr.3     3
#> 10     4 lyr.1     1
#> # ... with 290 more rows

^{Created on 2022-07-12 by the reprex package (v2.0.1)}

CodePudding user response：

With

library(terra)
y <- rast(ncol=10, nrow=10, nlyr=3, vals=1:300)

You can get the values ordered as you want them like this

v <- values(y)
x <- as.vector( t(v) )

The trick is to use t (transpose) before using as.vector; otherwise the order of the values would be by layer (column in x) instead of by cell.

You can turn vector x into a matrix with

m <- matrix(x, ncol=1)
head(m)
#     [,1]
#[1,]    1
#[2,]  101
#[3,]  201
#[4,]    2
#[5,]  102
#[6,]  202