Grouping a data frame with dplyr and plyr within the same function using prequoted arguments

Question

I have a somewhat complicated function with many arguments. Without going into too many details - it does some calculations on a data.frame using dplyr and plyr and returns a data.frame with several results columns attached.

I have an argument for the main grouping variable and ... for any additional groups. I use these grouping variables both with dplyr and plyr.

The dplyr bit was pretty straightforward - I used enquo and !! without issue. But I cannot figure out how to use the same principles with plyr.

It works with dplyr.

myfun <- function(data, main_group, ...) {
  group <- enquo(main_group)
  add_groups <- enquos(...)
  data %>%
    group_by(!! group, !!! add_groups)
}

mydata <- data.frame(a = 1:3, b = 1:3, c = 1:3, d = 1:3)
myfun(mydata, main_group = a, b, c)

But not with plyr and I need both.

myfun <- function(data, main_group, ...) {
  group <- enquo(main_group)
  add_groups <- enquos(...)
  ddply(data, .(!! group, !!! add_groups), .fun = function(X) { data.frame(result1 = 1, result2 = 1, result3 = 1) })
}
myfun(mydata, main_group = a, b, c)

I suppose it would be easiest if I could use a vector of string variable names in ddply call.

ddply(mydata, c("a", "b", "c"), .fun = function(X) { data.frame(result1 = 1, result2 = 1, result3 = 1) })

But how can I get c("a", "b", "c") within the function if the argument values are a, b, c?

rlang functions work in dplyr and not in plyr. Also check the githug page where it states that plyr is retired — akrun
OK, but if you have a variable name as a string (say "a"), but the function uses unquoted names (a), you can use sym("a") to get a. Is there something that does the reverse so I could use the resulting strings in plyr? — M4RT1NK4

Mikael Poul Johannesson Mikael Poul Johannesson · Accepted Answer · 2019-10-24T17:19:18

Please do not use plyr, it is retired and outdated. For similar functionality, please see purrr.

However, I do believe you can use deparse(substitute()) to solve your problem. E.g.,

myfun <- function(data, main_group, ...) {
  group <- deparse(substitute(main_group))
  add_groups <- as.character(as.list(match.call(expand.dots = FALSE)$...))
  ddply(data, c(group, add_groups), .fun = function(X) { data.frame(result1 = 1, result2 = 1, result3 = 1) })
}

Grouping a data frame with dplyr and plyr within the same function using prequoted arguments

1 Answers