如何使用相当于 R 中 foreach 循环的 Stata 和 mutate()

Question

我的原始数据中有 4 个不同的变量（全部以前缀 raw_ 开头），后面是时间/金钱和学校/社交的组合。在我的真实数据中，我对变量的每个值都有更多选项，因此最终会变成相当长的代码块，必须迭代 4 次。

必须有一种简单的方法可以在 R 中执行相当于 Stata 中的 foreach 循环（我更常使用它来进行数据清理），以将 4 次迭代压缩为 1 次。

在 Stata 中，循环看起来像这样：

foreach suffix in hours_school hours_social money_school money_social {
   gen olderch_`suffix' = .
   replace olderch_`suffix' = 90 if raw_`suffix' == "90% or more 10% or less"
   replace olderch_`suffix' = 10 if raw_`suffix' == "10% or less 90% or more"
}

这是我当前版本的 R 代码，我希望对其进行类似的压缩：

d <- d %>% 
  mutate(olderch_hours_school = case_when (
    raw_hours_school == "90% or more 10% or less" ~ 90,
    raw_hours_school == "10% or less 90% or more" ~ 10
  ))

d <- d %>% 
  mutate(olderch_hours_social = case_when (
    raw_hours_social == "90% or more 10% or less" ~ 90,
    raw_hours_social == "10% or less 90% or more" ~ 10
  ))

d <- d %>% 
  mutate(olderch_money_school = case_when (
    raw_money_school == "90% or more 10% or less" ~ 90,
    raw_money_school == "10% or less 90% or more" ~ 10
  ))

d <- d %>% 
  mutate(olderch_money_social = case_when (
    raw_money_social == "90% or more 10% or less" ~ 90,
    raw_money_social == "10% or less 90% or more" ~ 10
  ))

Answer 1

dplyr::across()

(docs) 允许您在数据框中的多个列上应用函数。您可以通过添加

case_when

来创建 lambda 函数并使用

~

代表每一列来调整您的

.

语句。

您可以提供所有正在转换的列的向量，或使用选择助手，例如

everything()

。

玩具数据遵循示例。

library(dplyr)


glimpse(df)
#> Rows: 10
#> Columns: 4
#> $ hours_school <chr> "90% or more 10% or less", "90% or more 10% or less", "90…
#> $ hours_social <chr> "10% or less 90% or more", "10% or less 90% or more", "10…
#> $ money_school <chr> "90% or more 10% or less", "10% or less 90% or more", "90…
#> $ money_social <chr> "90% or more 10% or less", "10% or less 90% or more", "90…

df |>
  mutate(across(
    c(hours_school, hours_social, money_school, money_social),
    ~ case_when(. ==  "90% or more 10% or less" ~ 90,
                . == "10% or less 90% or more" ~ 10)))
#>    hours_school hours_social money_school money_social
#> 1            90           10           90           90
#> 2            90           10           10           10
#> 3            90           10           90           90
#> 4            10           90           90           10
#> 5            90           10           90           10
#> 6            10           90           90           90
#> 7            10           10           10           90
#> 8            10           90           10           90
#> 9            90           90           90           90
#> 10           90           90           10           10

df |>
  mutate(across(
   everything(),
    ~ case_when(. ==  "90% or more 10% or less" ~ 90,
                . == "10% or less 90% or more" ~ 10)))
#>    hours_school hours_social money_school money_social
#> 1            90           10           90           90
#> 2            90           10           10           10
#> 3            90           10           90           90
#> 4            10           90           90           10
#> 5            90           10           90           10
#> 6            10           90           90           90
#> 7            10           10           10           90
#> 8            10           90           10           90
#> 9            90           90           90           90
#> 10           90           90           10           10

数据

set.seed(123)

df <- data.frame(
  hours_school = sample(c("90% or more 10% or less",
                          "10% or less 90% or more"),
                        10, replace = TRUE),
  hours_social = sample(c("90% or more 10% or less",
                          "10% or less 90% or more"),
                        10, replace = TRUE),
  money_school = sample(c("90% or more 10% or less",
                          "10% or less 90% or more"),
                        10, replace = TRUE),
  money_social = sample(c("90% or more 10% or less",
                          "10% or less 90% or more"),
                        10, replace = TRUE)
)