r - purrr - 将数据帧的行提取为向量

Question

我有一个数据框，希望将每一行提取到一个列表中 强制作为向量。我只想使用 tidyverse 和 purrr 包来实现这一点。

我尝试执行此操作时有以下reprex:

library(magrittr)
library(tidyverse)
# Create the raw dummy data frame
df <- data.frame(
    x = c("apple", "banana", "cherry"),
    pattern = c("p", "n", "h"),
    replacement = c("x", "f", "q"),
    stringsAsFactors = FALSE
)

# Define the function to extract the specific row index
# of the dataframe as a vector
get_row_vec <- function(df, row_idx){
    df %>% 
        dplyr::slice(row_idx) %>% 
        base::unlist() %>% 
        base::as.vector()
}

# Try and apply get_row_vec rowwise on the dataframe
# NOTE: This does not work! Need help to fix this
purrr::pmap(.l = df, ~get_row_vec(df  = .l, row_idx = 1))
#> Error in eval(lhs, parent, parent): object '.l' not found

有人可以帮忙纠正上面的代码吗？ 帮助我了解如何使用 purrr 执行此操作？

编辑:根据下面的评论，这是我通过 purrr 寻求的理想输出

# MANUAL version of desired output
output <- list(get_row_vec(df, 1),
               get_row_vec(df, 2),
               get_row_vec(df, 3))
output
#> [[1]]
#> [1] "apple" "p"     "x"    
#> 
#> [[2]]
#> [1] "banana" "n"      "f"     
#> 
#> [[3]]
#> [1] "cherry" "h"      "q"

谢谢

Answer 1

您可以使用purrr::transpose来实现此目的:

library(purrr)
map(transpose(df), unlist, use.names = F)

#[[1]]
#[1] "apple" "p"     "x"    

#[[2]]
#[1] "banana" "n"      "f"     

#[[3]]
#[1] "cherry" "h"      "q"  

或者如果使用pmap:

pmap(df, c, use.names = F)

#[[1]]
#[1] "apple" "p"     "x"    

#[[2]]
#[1] "banana" "n"      "f"     

#[[3]]
#[1] "cherry" "h"      "q"