主页 / coding / 问题 / 78583412
Accepted
ksinva
Asked: 2024-06-06 04:45:40 +0800 CST

如何使用 mutate(across()) 中的字符串引用 ifelse() 函数的参数

  • 772

假设我有这个示例数据：

library(tidyverse)

colNames <- c('A', 'B')

df <- tibble(A = sample(1:1000, 20),
             A_Direction = 'Decreasing',
             B = sample(1:1000, 20),
             B_Direction = 'Increasing')

我想更改列A，B如果A_DirectionorB_Direction减少，则为 0，如果增加，则为 1。这是我尝试过的：

dfNew <- df %>%
  mutate(across(.cols = matches(paste0('^', colNames, '$')),
                .fns = ~ifelse(test = paste0(cur_column(), '_Direction')=='Decreasing',
                               yes = 0,
                               no = 1)))

但当然，它无法识别该test参数，因为它是一个字符串，因此两列都变成了1。我有什么想法可以通过编程来完成这个任务（OG 数据有很多很多这样的列）。

3 个回答

  1. Best Answer

    另一种选择是将列名称包装paste0(cur_column(), "_Direction")在 中get。

    注意：我还删除了ifelse.

    library(dplyr, warn = FALSE)

set.seed(123)

df %>%
  mutate(across(
    matches(paste0("^", colNames, "$")),
    ~ +(!get(paste0(cur_column(), "_Direction")) == "Decreasing")
  ))
#> # A tibble: 20 × 4
#>        A A_Direction     B B_Direction
#>    <int> <chr>       <int> <chr>      
#>  1     0 Decreasing      1 Increasing 
#>  2     0 Decreasing      1 Increasing 
#>  3     0 Decreasing      1 Increasing 
#>  4     0 Decreasing      1 Increasing 
#>  5     0 Decreasing      1 Increasing 
#>  6     0 Decreasing      1 Increasing 
#>  7     0 Decreasing      1 Increasing 
#>  8     0 Decreasing      1 Increasing 
#>  9     0 Decreasing      1 Increasing 
#> 10     0 Decreasing      1 Increasing 
#> 11     0 Decreasing      1 Increasing 
#> 12     0 Decreasing      1 Increasing 
#> 13     0 Decreasing      1 Increasing 
#> 14     0 Decreasing      1 Increasing 
#> 15     0 Decreasing      1 Increasing 
#> 16     0 Decreasing      1 Increasing 
#> 17     0 Decreasing      1 Increasing 
#> 18     0 Decreasing      1 Increasing 
#> 19     0 Decreasing      1 Increasing 
#> 20     0 Decreasing      1 Increasing

    或使用pick()：

    df %>%
  mutate(across(
    matches(paste0("^", colNames, "$")),
    ~ +(!pick(paste0(cur_column(), "_Direction")) == "Decreasing")[[1]]
  ))
#> # A tibble: 20 × 4
#>        A A_Direction     B B_Direction
#>    <int> <chr>       <int> <chr>      
#>  1     0 Decreasing      1 Increasing 
#>  2     0 Decreasing      1 Increasing 
#>  3     0 Decreasing      1 Increasing 
#>  4     0 Decreasing      1 Increasing 
#>  5     0 Decreasing      1 Increasing 
#>  6     0 Decreasing      1 Increasing 
#>  7     0 Decreasing      1 Increasing 
#>  8     0 Decreasing      1 Increasing 
#>  9     0 Decreasing      1 Increasing 
#> 10     0 Decreasing      1 Increasing 
#> 11     0 Decreasing      1 Increasing 
#> 12     0 Decreasing      1 Increasing 
#> 13     0 Decreasing      1 Increasing 
#> 14     0 Decreasing      1 Increasing 
#> 15     0 Decreasing      1 Increasing 
#> 16     0 Decreasing      1 Increasing 
#> 17     0 Decreasing      1 Increasing 
#> 18     0 Decreasing      1 Increasing 
#> 19     0 Decreasing      1 Increasing 
#> 20     0 Decreasing      1 Increasing
    • 3

  2. 如果你已经有了，colNames你可以进行命名向量查找：

    library(dplyr)
library(stringr)

lookup <- setNames(str_c(colNames, "_Direction"), colNames)

df |>
  mutate(across(all_of(colNames), \(x) as.numeric(pick(all_of(lookup[[cur_column()]]))[[1]] == "Increasing")))

    或者您可以像这样访问across()：

    df |>
  mutate(across(all_of(colNames), \(x) as.numeric(pick(all_of(str_glue("{cur_column()}_Direction")))[[1]] == "Increasing")))

    最后，如果您的真实数据具有更多值而不仅仅是增加/减少，您可以case_match()像这样使用，它也有一个default参数：

    df |>
  mutate(across(all_of(colNames), \(x) {
    case_match(
      pick(all_of(str_glue("{cur_column()}_Direction")))[[1]],
      "Increasing" ~ 1,
      "Decreasing" ~ 0
    )
  }))

    只需添加更多值，如图所示（例如"Constant" ~ NA）。

    pick()使用（而不是像df我在帖子评论中提到的那样）的优点是，如果您修改了管道中的变量，则此函数将检索当前状态的值：

    # Output truncated with `slice()` 
df |> 
  mutate(A_Direction = replace(A_Direction, 1, "Increasing")) |>
  mutate(across(all_of(colNames), \(x) as.numeric(pick(all_of(str_glue("{cur_column()}_Direction")))[[1]] == "Increasing"))) |>
  slice(1:3)
#       A A_Direction     B B_Direction
#   <dbl> <chr>       <dbl> <chr>      
# 1     1 Increasing      1 Increasing # notice A is correct value
# 2     0 Decreasing      1 Increasing 
# 3     0 Decreasing      1 Increasing
    • 2

  3. 这是一个实用的解决方法：

    library(dplyr)

df |> 
  select(1:2) |> 
  mutate(A = ifelse(A_Direction == "Decreasing", 0,1)) |> 
  bind_cols(df |> 
              select(3:4) |> 
              mutate(B = ifelse(B_Direction == "Decreasing", 0,1))
            )

    A A_Direction     B B_Direction
   <dbl> <chr>       <dbl> <chr>      
 1     0 Decreasing      0 Decreasing 
 2     0 Decreasing      0 Decreasing 
 3     0 Decreasing      0 Decreasing 
 4     0 Decreasing      0 Decreasing 
 5     0 Decreasing      1 Increasing 
 6     0 Decreasing      1 Increasing 
 7     0 Decreasing      0 Decreasing 
 8     1 Increasing      0 Decreasing 
 9     1 Increasing      0 Decreasing 
10     0 Decreasing      1 Increasing 
11     0 Decreasing      1 Increasing 
12     0 Decreasing      0 Decreasing 
13     1 Increasing      0 Decreasing 
14     0 Decreasing      1 Increasing 
15     1 Increasing      1 Increasing 
16     1 Increasing      1 Increasing 
17     0 Decreasing      1 Increasing 
18     0 Decreasing      1 Increasing 
19     0 Decreasing      1 Increasing 
20     0 Decreasing      0 Decreasing
    • 0

相关问题