Accessing to names of the elements processed by the R *apply functions

Question

When using the R *apply functions, sometime one would like to access to the names of the processed elements inside the apply processing function. To make it clear, consider first this for-loop:

df=data.frame(x=1:3, y=11:13)
for(i in 1:ncol(df)){
    cat('processing col', names(df)[i], '\n')
    print(df[[i]])
}        

# processing col x  
# [1] 1 2 3         
# processing col y  
# [1] 11 12 13      

When moving to the apply version:

lapply(df, function(col){
    ## processing col ??
    col})

# $x             
# [1] 1 2 3      
#                        
# $y             
# [1] 11 12 13

how to access the column names from inside the anonymous apply function?

Answer 1

This can't be done, except how you've done it with for. If you want to return a structure with the results (and would be appending to a list, for example), *apply may still be appropriate.

x <- lapply(names(iris), function(name) {
        cat("processing col", name, "\n")
        iris[[name]]
     })
## processing col Sepal.Length 
## processing col Sepal.Width 
## processing col Petal.Length 
## processing col Petal.Width 
## processing col Species 

names(x) <- names(iris)

x is now a list containing the columns of the data frame iris.

Answer 2

Two things to note here: First: The for loop, doesn't have a return value, but the lapply does. Second: The lapply does not pass the column name, only the unnamed vector.

So something like this a better generic solution:

x <- lapply(1:ncol(df),function(i){
  cat('processing col', names(df)[i], '\n')
  print(df[[i]])
  })

You can later discard x (by rm(x)) if you don't need it.

Note that this will be easier if you have to replace a lot of for loops!