Combining data based on user input in R

Question

I am attempting to write a function that combines data based on a user input.

I've searched around but cannot find anything that seems to help (I am somewhat new to using R).

Running each line separately works great. However, when I wrap them into a used defined function nothing is produced. For example, I have the following data: ASW1_2016, ASW_2017, ASW_2018, CSW_2015, CSW_2017. I want to prompt the user to input a site name "ASW1" and have the function combine all data that contain this name.

#Running these two lines separately works:
Site <- invisible(readline(prompt = "Enter Site Name:"))

assign(Site, rbindlist(mget(apropos(Site), inherits = TRUE))

#Putting these two lines into a function does not produce anything:
CombineData <- function()

{

Site <- invisible(readline(prompt = "Enter Site Name:"))

assign(Site, rbindlist(mget(apropos(Site), inherits = TRUE))

}

CombineData()

After being asked to enter a site name (e.g. ASW1) I expect all data containing this name to be combined and assigned the name provided by the user. However, as stated above, when I run CombineData() and enter the site name following the prompt, nothing happens.

Answer 1

assign called within a function by default does the assignment in the function's environment, instead of in the global environment. So basically, it's like storing the result in a variable within the function. Once the function returns, that value is lost. The following illustrates this (this is essentially what you are doing):

fxn <- function() assign("xyz", 1)
fxn()
print(xyz)
# Error in print(xyz) : object 'xyz' not found

If you want to use assign to store in the global environment (outside of the function), you could add argument envir = .GlobalEnv to your call to assign.

fxn <- function() assign("xyz", 1, envir = .GlobalEnv)
fxn()
print(xyz)
# [1] 1

That being said, using assign to store values is often not the preferred way to store data in R (you can read more about that at Why is using assign bad?). You might find it easier to maintain your code if you stored it in a list. As an example, if you had an initial list l <- list(), then you could store your new data in that list with l[[Site]] <- rbindlist(mget(apropos(Site), inherits = TRUE) and access it with [[ or $.