将元素结构可变的嵌套列表展平为带ID与分段标识的DataFrame

Hey there! I get it—dealing with nested lists that have mixed structures (some standalone 2D arrays, some sub-lists of arrays) can be a total pain, especially when plyr::ldply chokes on the inconsistencies. Let's use data.table to build a clean, concise solution that handles both cases smoothly.

Step 1: Example Data #

First, let's replicate your nested list structure so we can test the solution:

# Sample nested list with mixed structures
gc_all <- list(
  # Parent element 1: 2D array (lon x lat)
  array(rnorm(4), dim = c(2, 2), dimnames = list(c("lon1", "lon2"), c("lat1", "lat2"))),
  # Parent element 2: Sub-list containing two 2D arrays
  list(
    array(rnorm(4), dim = c(2, 2), dimnames = list(c("lon1", "lon2"), c("lat1", "lat2"))),
    array(rnorm(4), dim = c(2, 2), dimnames = list(c("lon1", "lon2"), c("lat1", "lat2")))
  ),
  # Parent element 3: Another standalone 2D array
  array(rnorm(4), dim = c(2, 2), dimnames = list(c("lon1", "lon2"), c("lat1", "lat2")))
)

Step 2: The Concise data.table Solution #

Here's a near-one-liner that flattens the list while preserving id (parent element index) and section (sub-element index, or 1 for standalone arrays):

library(data.table)

flattened_dt <- rbindlist(
  lapply(seq_along(gc_all), function(id) {
    current_elem <- gc_all[[id]]
    # Normalize: wrap standalone arrays into a single-element list to match sub-list structure
    elem_list <- if (is.list(current_elem)) current_elem else list(current_elem)
    
    # Process each sub-element/array into a data.table with id and section
    rbindlist(
      lapply(seq_along(elem_list), function(section) {
        as.data.table(as.data.frame.table(elem_list[[section]], responseName = "value"))[,
          `:=`(id = id, section = section)
        ]
      })
    )
  })
)[, c("id", "section", "Var1", "Var2", "value")] # Reorder columns for clarity

How This Works: #

• seq_along(gc_all) generates the id for each parent list element
• We normalize the input: if an element is a standalone array, we wrap it into a list so we can process it the same way as sub-lists
• as.data.frame.table() converts each 2D array into a long-format frame with Var1 (lon), Var2 (lat), and value
• We add the id and section columns using data.table's efficient := operator
• rbindlist() seamlessly combines all the individual data.tables into one final table

Step 3: Expected Output #

When you run the code, you'll get a clean data.table (convertible to a data frame with as.data.frame()) that looks like this:

head(flattened_dt)
#    id section Var1 Var2      value
# 1:  1       1 lon1 lat1 -0.1234567
# 2:  1       1 lon2 lat1  0.9876543
# 3:  1       1 lon1 lat2 -0.5678901
# 4:  1       1 lon2 lat2  0.2345678
# 5:  2       1 lon1 lat1  1.2345678
# 6:  2       1 lon2 lat1 -0.8765432

This approach is way more concise than nested lapply + ifelse chains, and data.table's optimized operations make it fast even for large lists.

内容的提问来源于stack exchange，提问作者jogall