create names of loop when iterating data frame

Question

mylist <- list(demographic_vars, usage_engagement_vars, billing_contracttype_vars, contract_type_vars)

Each variable in mylist is a dataframe.

I'm writing a function to pass to lapply(). I would like to loop over mylist and name a variable like so:

varData <- function(x) {
    paste('cv_prediction',x,sep='') <- data.frame()
}

Goal of the above would be this: (e.g. for the first iteration)

cv_predictiondemographic_vars <- data.frame()

I'm getting an error in my function that I suspect is related to the above.

Error in paste("cv_prediction", x, sep = "") <- data.frame() :
target of assignment expands to non-language object

If I want to create variables within my function that are the concatenation of the current value of x in a loop to some other string, how would I do that?

I guess you need assign if you are creating a data.frame object — akrun
– akrun, Commented Nov 23, 2015 at 12:59
Thanks but more generally how does one use the iteration name in R? E.g. in Javascript more often than not the iterator i is a number so somevar + i = something does the trick. Would I have to use paste() to mirror this in R? What is the "conventional" R way of naming vars within a loop? — Doug Fir
– Doug Fir, Commented Nov 23, 2015 at 13:06

Konrad Rudolph · Accepted Answer · 2015-11-23 13:50:03Z

It’s tricky. In general, you’d write the function like this:

varData = function(x) {
    name = paste0('cv_prediction', deparse(substitute(x), backtick = TRUE))
    assign(name, x, parent.frame())
}

This works when called like this, say:

varData(some_data_frame)

However, it won’t work with a loop variable since then it will use the name of that variable. As far as I know there’s no way of actually achieving what you want. Instead, you need to replace list in your above code by something else:

named_list = function (...) {
    names = sapply(match.call()[-1], deparse, backtick = TRUE)
    setNames(list(...), names)
}

And then use it:

mylist = named_list(demographic_vars, usage_engagement_vars, billing_contracttype_vars, contract_type_vars)

This will result in a mylist whose values are named by the objects. For example:

a = 1
b = 2
named_list(a, b)
# $a
# [1] 1
#
# $b
# [1] 2

Now you can start using the names of the list in your code.

Vongo · Accepted Answer · 2015-11-23 13:28:36Z

1

You could consider using eval (see doc).

> varData <- function(x) {
+    eval(parse(text=paste('cv_prediction', x,' <<- data.frame()',sep='')))
+ }
> varData(3)
> ls()
[1] "cv_prediction3" "varData"
> class(cv_prediction3)
[1] "data.frame"

Note that instead of using <<- operator you can direcly address the environment where you want your new object to be created :

eval(parse(text=paste('cv_prediction', x,' <- data.frame()',sep='')),env=.GlobalEnv)

edited Nov 23, 2015 at 13:28

answered Nov 23, 2015 at 13:24

Vongo

1,4441 gold badge21 silver badges30 bronze badges

1 Comment

Konrad Rudolph Over a year ago

While this works, eval is really not the right solution here — assign is (lower conceptual complexity, vastly more efficient, etc.).

Collectives™ on Stack Overflow

create names of loop when iterating data frame

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related