R Edit data frame in function within function - r

Question

I have a code made up of a lot of functions used for different codes and which will modify a df by adding some columns. I need to have a global function that takes over several of these functions, but since they are functions inside another function, my df does not update this on every function call. Do you have any advice for this problem?
Here is an example of my problem :
f_a<-function(df){
df$x<-1
.GlobalEnv$df <- df
}
f_b<-function(df){
df$y<-1
.GlobalEnv$df <- df
}
f_global<-function(df){
f_a(df)
f_b(df)
}
In this case df will not have the x and y columns created
Thanks

Answer 1

It's generally a bad idea for functions to have "side effects": things are easier to get right if functions are completely self contained. For your example, that would look like this:
f_a<-function(df){
df$x<-1 # This only changes the local copy
df # This returns the local copy as the function result
}
f_b<-function(df){
df$y<-1
df
}
f_global<-function(df){
df <- f_a(df) # This uses f_a to change the local copy
df <- f_b(df) # This uses f_b to make another change
df # This returns the changed dataframe
}
Then you use it like this:
mydf <- data.frame(z = 1)
mydf <- f_global(mydf)

Answer 2

use this operator <<- in the function.as an example:
dat = data.frame(x1 = rep(1,10),x2 = rep(2,10),x3 = rep(3,10))
head(dat)
myFun <- function(x){
print(x)
dat$x1 <<- rep(5,10)
}
myFun(10)
head(dat)

Answer 3

In the call to f_b the input argument df is assigned to .GlobalEnv rewriting the df that already existed there. So f_global first calls f_a and creates a column x, then calls f_b passing it its input data.frame and f_b creates a column y in this df.
All that needs to be changed is f_global:
f_global<-function(df){
f_a(df)
f_b(.GlobalEnv$df)
}
f_global(data.frame(a=1))
df
# a x y
#1 1 1 1
df <- head(mtcars)
f_global(df)
df
# mpg cyl disp hp drat wt qsec vs am gear carb x y
#Mazda RX4 21.0 6 160 110 3.90 2.620 16.46 0 1 4 4 1 1
#Mazda RX4 Wag 21.0 6 160 110 3.90 2.875 17.02 0 1 4 4 1 1
#Datsun 710 22.8 4 108 93 3.85 2.320 18.61 1 1 4 1 1 1
#Hornet 4 Drive 21.4 6 258 110 3.08 3.215 19.44 1 0 3 1 1 1
#Hornet Sportabout 18.7 8 360 175 3.15 3.440 17.02 0 0 3 2 1 1
#Valiant 18.1 6 225 105 2.76 3.460 20.22 1 0 3 1 1 1
Though the code above works and follows the lines of the question, I think that a better strategy is to have f_global change its input argument assigning the return value of each f_* and assign the end result in f_global's parent environment only after all transformations are done.
f_a <- function(df){
df$x <- 1
df
}
f_b <- function(df){
df$y <- 1
df
}
f_global <- function(df){
dfname <- deparse(substitute(df))
df <- f_a(df)
df <- f_b(df)
assign(dfname, df, envir = parent.frame())
invisible(NULL)
}
df1 <- data.frame(a=1)
f_global(df1)
df1
df <- head(mtcars)
f_global(df)
df