CODEWITHSUNDEEP
rdplyr
ANam
ANam

Reputation: 347

How to assign a name to an output global dataframe in a function?

I have about 12 datasets that I am trying to clean before implementing them as datatables in my final code. I currently have code that works for individual datasets but I want to implement this as a function so that I don't have to re-type the code blocks over and over. 

countyshapesClean <- function(inputdata, outputdata) {
  outputdata <<- inputdata %>%
select(
  county_cl,
  state_abb,
  avg_opioid,
  avg_oxy_hydro,
  avg_opioid_perc,
  avg_oxy_hydro_perc,
  avg_opioid_ppp,
  avg_oxy_hydro_ppp
) %>% 
mutate(
  avg_opioid = format(round(as.numeric(avg_opioid),2), nsmall = 2, big.mark = ","),
  avg_oxy_hydro = format(round(as.numeric(avg_oxy_hydro),2), nsmall = 2, big.mark = ","),
  avg_opioid_perc = format(round(as.numeric(avg_opioid_perc),2), nsmall = 2, big.mark = ","),
  avg_oxy_hydro_perc = format(round(as.numeric(avg_oxy_hydro_perc),2), nsmall = 2, big.mark = ","),
  avg_opioid_ppp = format(round(as.numeric(avg_opioid_ppp),2), nsmall = 2, big.mark = ","),
  avg_oxy_hydro_ppp = format(round(as.numeric(avg_oxy_hydro_ppp),2), nsmall = 2, big.mark = ",")
)
}

countyshapesClean(inputdata = countymerge2006, outputdata = countymerge2006clean)

This works but my output dataset is named outputdata instead of countymerge2006clean. How would I change this code to get my desired result?

Upvotes: 1

Views: 42

Answers (1)

MrFlick
MrFlick

Reputation: 206197

Don't pass in outputdata to the function. Instead, do something like

countymerge2006clean <- countyshapesClean(inputdata = countymerge2006)

where your function returns the updated data.frame.

countyshapesClean <- function(inputdata) {
  inputdata %>%
  select(...) %>% 
  mutate(...)
}

Proper functions in R should not create variables outside their own scope. (It's not considered good practice to use <<-).

Of course, almost anything is possible is R, and you could create variables outside the function scope with the assign() function (though I strongly recommend you do NOT do this). That would look something like this

countyshapesClean <- function(inputdata, outputdata) {
  outvar <- rlang::ensym(outputdata)
  new <- inputdata %>%
    select(...) %>% 
    mutate(...)
  assign(rlang::quo_name(outvar), new, parent.frame())
}
countyshapesClean(inputdata = countymerge2006, outputdata = countymerge2006clean)

Upvotes: 2

Related Questions