使用purrr执行逐行操作时遇原子向量$访问错误求助

Hey there! Let's break down why you're getting that "cannot use $ operator on atomic vector" error and fix it with purrr (or a few other straightforward approaches).

The Root of the Problem #

When you use apply(iris, 1, test_function), apply converts each row of your data frame into an atomic vector (not a single-row data frame object). Your function relies on the $ operator, which only works with lists or data frames—atomic vectors don't recognize this syntax, hence the error.

Fix 1: Use purrr to Pass Rows as Data Frames #

Since you wanted to use purrr anyway, this is the cleanest, most idiomatic approach. We’ll split the iris data frame into a list of single-row data frames, then use map_df to apply your function and automatically bind the results into a single data frame:

library(purrr)
library(dplyr)

test_function = function(dat_) {
  petal_width = dat_$Petal.Width
  sepal_width = dat_$Sepal.Width
  petal_length = dat_$Petal.Length
  sepal_length = dat_$Sepal.Length
  
  tibble(petal_length, sepal_length, sepal_width, petal_width)
}

# Split iris into list of single-row data frames, then apply function
result <- iris %>% 
  split(seq(nrow(.))) %>% 
  map_df(test_function)

Fix 2: Adjust Your Function to Work with Atomic Vectors #

If you still want to use apply, modify your function to use [[]] instead of $—this operator works with both data frames and named atomic vectors (which apply produces here):

test_function = function(dat_) {
  # Use [[]] to handle both data frames and atomic vectors
  petal_width = dat_[["Petal.Width"]]
  sepal_width = dat_[["Sepal.Width"]]
  petal_length = dat_[["Petal.Length"]]
  sepal_length = dat_[["Sepal.Length"]]
  
  tibble(petal_length, sepal_length, sepal_width, petal_width)
}

# Now apply runs without errors
result <- apply(iris, 1, test_function) %>% bind_cols()

Fix 3: Convert Vectors to Data Frames in apply #

A quick workaround is to convert each atomic vector back to a data frame before passing it to your function:

result <- apply(iris, 1, function(row) test_function(as.data.frame(t(row)))) %>% 
  bind_cols()

A Quick Simplification #

You can also streamline your function to avoid repetitive column extractions—since all you’re doing is rearranging columns, this shorter version works just as well:

test_function = function(dat_) {
  dat_ %>% select(Petal.Length, Sepal.Length, Sepal.Width, Petal.Width)
}

This works seamlessly with both single-row data frames (from purrr) and converted vectors (from the other fixes).

内容的提问来源于stack exchange，提问作者spazznolo