Transform multiple columns with a function that uses different arguments per column

Question

I write and review a fair amount of R code like this:

df <- data.frame(replicate(10, sample(0:5, 10, rep = TRUE)))
my.func <- function(col, y) {col %in% y}

df$X2 <- my.func(df$X2, c(1,2))
df$X3 <- my.func(df$X3, c(4,5))
df$X5 <- my.func(df$X5, c(1,2))
df$X6 <- my.func(df$X6, c(4,5))
df$X8 <- my.func(df$X8, c(4,5))
df$X9 <- my.func(df$X9, c(1,2))
df$X10 <- my.func(df$X10, c(1))

That is, certain columns in a data.frame (or data.table) are transformed using a function, where one argument is a column and the other is some arbitrary, somewhat-unique-to-that-column value.

What's a more concise way to make such transformations?

I've tried using data.table's set (:=) operator, which makes things slightly cleaner, but still each column name must appear twice and the function must appear once for each column.

Answer 1

A concise way would be Map with the input arguments as the dataset ('df') and a list of vector that would be passed as argument to my.func. Here, each column of the data.frame is a unit and similarly the vector element from list.

df[] <- Map(my.func, df, list(1:2, 4:5, 3:4))

NOTE: The OP's function or a minimal reproducible example is not provided, so it is not tested

NOTE2: Here, the assumption is that the number of columns is 3. If it is more than 3, increase the length of the list as well

---

The above can also be converted to data.table syntax

library(data.table)
setDT(df)[, names(df) := Map(my.func, .SD, list(1:2, 4:5, 3:4))]

If only a subset of columns needs to be changed, specify the columns in .SDcols, and also change the names(df) to the subset of names

---

Or with tidyverse

library(tidyverse)
map2_dfc(df, list(1:2, 4:5, 3:4), my.func)

Answer 2

OP's request from a comment:

make the association between column names and function argument(s) for those columns more explicit

Adjusting the Map approach seen in the other answers:

yL <- list(X2 = 1:2, X3 = 4:5, X5 = 3:4, X6 = 4:5, X8 = 4:5, X9 = 1:2, X10 = 1)
df[names(yL)] <-  Map(my.func, df[names(yL)], y = yL)

With data.table:

# this saves you from writing DT twice
DT[, names(yL) := Map(my.func, .SD, y = yL), .SDcols=names(yL)]