Change the names of multiple cells in R using a function - r

Question

Consider the following data frame:
df <- setNames(data.frame(1:5,rep(1,5)), c("id", "value"))
I want to change the names for multiple cells in the column "id". Let's say I want to change the following:
df$id[df$id %In% 2:3] <- 1
df$id[df$id == 4] <- 3
However, instead of using the code above, I want to create a function, where I can do the transformation more "smooth" (because I have a lot of data frames, where I need to change the names for the cells). I want to create a function:
mapping <- function(...) {
...
}
where I afterward can create a simple and smooth mapping function for my df, where I only have to specific the "old" and the "new" names for the cells. Something like this:
df_mapping <- function(...) {
2.1
3.1
4.3
}
And then I can apply the function on my data and specific which column it should do it for, and it will work in the same way as the code with gsub:
df <- df_mapping(df,id)
Is it possible to create that mapping function?

Answer 1

if we need a function, then can have a 'data' argument, column name, values to replace and replacer value, then create the logical condition, subset the column, assign with replacer_val and return the dataset after the assignment
f1 <- function(dat, colnm, values_to_replace, replacer_val) {
dat[[colnm]][dat[[colnm]] %in% values_to_replace] <- replacer_val
return(dat)
}
f1(df, "id", c(2, 3), 1)
-output
# id value
#1 1 1
#2 1 1
#3 1 1
#4 4 1
#5 5 1
To replace values with corresponding sets of replacers,
f2 <- function(dat, colnm, values_to_replace, replacer_vals) {
nm1 <- setNames(replacer_vals, values_to_replace)
v1 <- nm1[as.character(dat[[colnm]])]
i1 <- !is.na(v1)
dat[[colnm]][i1] <- v1[i1]
return(dat)
}
f2(df, "id", c(2, 3), c(5, 6))
# id value
#1 1 1
#2 5 1
#3 6 1
#4 4 1
#5 5 1
Or another option is to create a key/value dataset and use merge or join
library(data.table)
f3 <- function(dat, colnm, values_to_replace, replacer_vals) {
keydat <- data.frame(key = values_to_replace, val = replacer_vals)
names(keydat)[1] <- colnm
dt <- as.data.table(dat)
dt[keydat, (colnm) := val, on = colnm][]
return(dt)
}
f3(df, "id", c(2, 5), c(3, 6))

Answer 2

Maybe a mapping like below could help
mapping <- function(df, id, to_replace, obj_value) {
transform(df, id = replace(id, id %in% to_replace, obj_value))
}
e.g.,
> mapping(df, id, c(2, 3), 1)
id value
1 1 1
2 1 1
3 1 1
4 4 1
5 5 1

Answer 3

You can use dplyr's recode function
mapping <- function(data, col, old, new) {
data[[col]] <- dplyr::recode(data[[col]], !!!setNames(new, old))
data
}
mapping(df, "id", c(2, 3), c(7L, 8L))
# id value
#1 1 1
#2 7 1
#3 8 1
#4 4 1
#5 5 1