如何使用purrr::map重组列表中RasterStack/RasterBrick的图层

Got it, let's walk through how to group your RasterBrick layers by their bio names using purrr. I'll use sample data to make this concrete, so you can follow along and adapt it to your own dataset.

Step 1: Setup Sample Data (Skip if you have your own) #

First, let's create a list of RasterBricks to work with. Each brick has layers named bio_1, bio_2, and bio_3:

library(raster)
library(purrr)

# Generate 3 sample RasterBricks
set.seed(123)
create_sample_brick <- function() {
  brick(nl = 3, nrow = 5, ncol = 5) %>%
    setNames(paste0("bio_", 1:3))
}

brick_list <- list(create_sample_brick(), create_sample_brick(), create_sample_brick())

Step 2: Extract Unique Bio Layer Names #

First, we need to get all the unique bio layer names we want to group by. This works even if your bricks have different sets of bio layers (though I assume they're consistent):

# Get sorted unique bio names across all bricks
unique_bio_names <- brick_list %>%
  map(names) %>%          # Get layer names from each brick
  flatten_chr() %>%       # Combine into a single character vector
  unique() %>%            # Keep only unique names
  sort()                  # Optional: ensure order is bio_1, bio_2, etc.

Step 3: Group Layers by Bio Name Using purrr #

Now, we'll use purrr::map to iterate over each unique bio name, extract that layer from every brick in your list, and stack them into a new RasterBrick:

# Group layers into new RasterBricks
grouped_bricks <- unique_bio_names %>%
  map(function(bio_name) {
    brick_list %>%
      map(~ .x[[bio_name]]) %>%  # Extract the bio_name layer from each brick
      brick()                    # Combine extracted layers into a new RasterBrick
  })

# Name the list elements for clarity (optional but helpful)
names(grouped_bricks) <- unique_bio_names

What this does: #

• For each bio name (like bio_1), we loop through your original list of bricks and pull out the bio_1 layer from each one.
• We then combine those individual layers into a single RasterBrick (use stack() instead of brick() if you prefer a RasterStack).

Step 4: Convert to Numeric Matrix (As You Mentioned) #

Once you have your grouped bricks, converting them to matrices is straightforward. You can use raster::values() or as.matrix():

# Convert each grouped brick to a matrix
bio_matrices <- grouped_bricks %>%
  map(values)  # Or map(as.matrix) depending on your needs

Notes #

• If some bricks are missing a specific bio layer, this will throw an error. If that's a possibility, you can add a check with purrr::possibly to skip missing layers or handle them gracefully.
• This approach keeps your code consistent with tidyverse/purrr style, avoiding explicit for loops and making it easy to extend if you add more bricks or bio layers later.

内容的提问来源于stack exchange，提问作者Javier Fajardo