r - 使用三个点/省略号 (...) 作为 dplyr 和 merge() 的列名

Question

我有一个函数，我首先使用 group_by()传递给函数的多个列名，然后合并这些相同的列名。

我可以做到 group_by()使用 vars <- enquos(…)其次是 group_by(!!!vars) .这要求传递给函数的列名不在引号中，否则 group_by()创建包含引号的新列名。

我可以使用 vars <- list(…) 进行合并其次是 merge(x, y, by = paste(vars)) .然而，这需要列名 在引号中，否则 list(…)生成:

error (object 'x' not found).

如何同时处理这两种情况？

例子:

example <- function(...) {
  vars1 <- enquos(...) #doesn't work unless (...) isn't in quotes
  #vars2 <- list(...) #doesn't work unless (...) is in quotes

  df1 <- mtcars %>%
    group_by(!!!vars1) %>%
    summarise(Mean = mean(mpg))

  df2 <- mtcars %>%
    group_by(!!!vars1) %>%
    summarise(Median = median(hp))

  merged <- merge(df1, df2, by = "cyl") #for functioning example only, want by to be (...)
  #merged <- merge(df1, df2, by = paste(vars2)) #doesn't work unless (...) is in quotes
  return(merged)
}

out <- example(cyl)

Answer 1

这是一个使用引用变量作为输入的版本:

example <- function(...) {
  vars1 <- syms(...) 
  vars2 <- list(...)
  df1 <- mtcars %>%
         group_by(!!!vars1) %>%
         summarise(Mean = mean(mpg))

  df2 <- mtcars %>%
          group_by(!!!vars1) %>%
          summarise(Median = median(hp))

 merged <- merge(df1, df2, by = vars2[[1]])
 return(merged)

}

example("cyl")
#  cyl  Mean Median
#1   4 26.66   91.0
#2   6 19.74  110.0
#3   8 15.10  192.5

example(c("cyl", "am"))
#  cyl am  Mean Median
#1   4  0 22.90   95.0
#2   4  1 28.07   78.5
#3   6  0 19.12  116.5
#4   6  1 20.57  110.0
#5   8  0 15.05  180.0
#6   8  1 15.40  299.5