如何用separate类函数对DataFrame所有列执行拆分操作？

问题场景 #

现有一个DataFrame，所有列的内容都遵循x_y格式，需要将每一列拆分为两列。单独对某一列使用separate_wider_delim可正常完成拆分，但结合mutate和across批量处理所有列时触发报错，提示data必须是数据框而非字符向量。

测试数据 #

test <- structure(list(A = c("511686_0.112", "503316_0.105", "476729_0.148", 
"229348_0.181", "385774_0.178", "209277_0.029", "299921_0.124", 
"486771_0.123", "524146_0.07", "496030_0.119"), B = c("363323_0.103", 
"260709_0.105", "361361_0.148", "731426_0.181", "222799_0.178", 
"140296_0.029", "388191_0.124", "500136_0.123", "487344_0.07", 
"267303_0.119"), C = c("362981_0.103", "260261_0.105", "360912_0.148", 
"730423_0.181", "222351_0.178", "139847_0.029", "379717_0.124", 
"499662_0.123", "486869_0.07", "266907_0.119")), class = c("tbl_df", 
"tbl", "data.frame"), row.names = c(NA, -10L))

单个列可行代码 #

test2 <- test %>%
  separate_wider_delim(A, delim = "_", names_sep = "_")

批量拆分报错代码 #

test3 <- test %>%
  mutate(across(everything(), separate_wider_delim, delim = "_", names_sep = "_"))

报错信息 #

Error in `mutate()`:
ℹ In argument: `across(everything(), separate_wider_delim, delim = "_", names_sep = "_")`.
Caused by error in `across()`:
! Can't compute column `A`.
Caused by error in `fn()`:
! `data` must be a data frame, not a character vector.
Run `rlang::last_error()` to see where the error occurred.

报错原因 #

separate_wider_delim是作用于整个数据框的函数，第一个参数要求传入数据框；而across会将每一列作为单独的字符向量传递给目标函数，导致参数类型不匹配，触发报错。

解决方案 #

方法一：直接在数据框层面批量拆分（推荐） #

无需使用mutate和across，直接调用separate_wider_delim并指定所有列即可：

test3 <- test %>%
  separate_wider_delim(everything(), delim = "_", names_sep = "_")

• everything()表示选中所有列进行拆分
• names_sep = "_"会让拆分后的新列名格式为原列名_1、原列名_2，例如原列A拆分为A_1和A_2

方法二：结合across的兼容写法（仅作参考） #

如果一定要用across，需先将单列向量转换为单列数据框，处理后再展开：

library(tidyr)

test3 <- test %>%
  mutate(across(everything(), ~ {
    # 将单列向量转为单列数据框
    tibble(col = .x) %>%
      separate_wider_delim(col, delim = "_", names_sep = "_")
  })) %>%
  # 展开嵌套的列
  unnest_wider(everything())

验证结果 #

执行推荐方法后，原DataFrame的每一列都会被拆分为两列，例如原列A的"511686_0.112"会被拆分为A_1列的511686和A_2列的0.112。

内容的提问来源于stack exchange，提问作者canderson156