R: How to subset DFs according to unique values in column and loop each of them in the function?

Question

I want to create a loop function, which subset data according to a unique value in Category column, so to have separated DFs where `Category == "A", Category == "B",.... Category == "F"

data <- structure(list(Category = c("A", "A", "B", "B", "C", "C", "C", 
"D", "E", "E", "E", "C", "A", "B", "B", "F", "F"), Sales = c(34L, 
32L, 34L, 56L, 32L, 66L, 6L, 55L, 66L, 56L, 3L, 43L, 56L, 65L, 
34L, 65L, 43L), Year = c(2015L, 2015L, 2015L, 2015L, 2016L, 2016L, 
2016L, 2016L, 2017L, 2017L, 2017L, 2017L, 2018L, 2018L, 2018L, 
2018L, 2018L)), class = "data.frame", row.names = c(NA, -17L))

   Category Sales Year
1         A    34 2015
2         A    32 2015
3         B    34 2015
4         B    56 2015
5         C    32 2016
6         C    66 2016
7         C     6 2016
8         D    55 2016
9         E    66 2017
10        E    56 2017
11        E     3 2017
12        C    43 2017
13        A    56 2018
14        B    65 2018
15        B    34 2018
16        F    65 2018
17        F    43 2018

and after throwing all these unique DFs into function, for example:

divide <- function(df) {
  Sales <- (Sales - 32) * 5 / 9
  return(df)
}

data <- divide(data)

and after looping all these DFs, combine them using rbind()

All these need to be automatized, I can't code and subset DFs manually for every unique value in the Category column. So I need to use unique() function

P.S This is not the original function I posted here! It is just an example. But the idea is to subset DFs with unique values in the Category column, throw them into the function and combine rbind() in the end

Answer 1

You can:

• Use split to split your data according to the unique Categories.
• Use lapply to apply a function to each of your dataset (here, I created a new variable, sumYear).
• Use do.call to get back to your original dataframe

divide <- function(var) {
  (var - 32) * 5 / 9
}

split(data, data$Category) |>
  lapply(function(x) transform(x, Sales = divide(x$Sales))) |>
  do.call(rbind.data.frame, args = _)

     Category      Sales Year
A.1         A   1.111111 2015
A.2         A   0.000000 2015
A.13        A  13.333333 2018
B.3         B   1.111111 2015
B.4         B  13.333333 2015
B.14        B  18.333333 2018
B.15        B   1.111111 2018
C.5         C   0.000000 2016
C.6         C  18.888889 2016
C.7         C -14.444444 2016
C.12        C   6.111111 2017
D           D  12.777778 2016
E.9         E  18.888889 2017
E.10        E  13.333333 2017
E.11        E -16.111111 2017
F.16        F  18.333333 2018
F.17        F   6.111111 2018

---

Note that this is not a very straightforward way to apply a function by group. You could use dplyr::group_by or ave in base R to do so; see e.g. solutions of this question: Run a custom function on a data frame in R, by group.