使函数返回多个对象时如何删除索引括号？

Question

I'm trying to use functions to cut down on lines of code, but my output isn't as clean as I'd like. Specifically, when returning multiple elements from a function (by making a list and using return ), R also prints what I think are indexing numbers (ex. [[1]][[1]][[1]] ). To make my markdown document cleaner and more readable, I'd like to remove those little numbers.

Any thoughts on how to do this? Looking around SO didn't really lead me anywhere. cat() doesn't work here since I'm trying to return a plot and a table.

An example using the iris dataset (with ggplot2, dplyr, and kableExtra packages):

data(iris)

TypeLoop <- function(fun) {
  all <- fun(c("setosa","versicolor","virginica"),"All")
  set <- fun("setosa","Setosa")
  vers <- fun("versicolor","Versicolor")
  virg <- fun("virginica","Verginica")
  out <- list(all,set,vers,virg)
  return(out)
}

PlotNTable <- function(type,name) {
  myplot <- ggplot(iris,aes(Sepal.Length)) +
    geom_histogram(data=filter(iris, Species %in% type),color="white",fill="darkblue") +
    labs(x="Sepal Length", y="Frequency",
         title=paste(name,": Sepal Length"))

  sep.table <- iris %>%
    filter(Species %in% type) %>%
    group_by(Sepal.Length) %>%
    summarize("Less than 6" = sum(Sepal.Length<6),
                "6 to 7" = sum(Sepal.Length<7 & Sepal.Length>=6),
                "More than 7" = sum(Sepal.Length>7))
  mytable <- kable(sep.table, caption = paste(name,": Sepal Length")) %>%
    kable_styling(c("striped","bordered","hover")) %>%
    column_spec(1, bold = T) 

  #Make the function return both objects
  out <- list(myplot,mytable)
  return(out)
}

#Return the plots and tables
TypeLoop(PlotNTable)

I include the results='asis' option in the top of the code chunk to make it actually knit the markdown document.

Also, I'm fairly new to R, so if you see red flags/redundancies/better ways to do things I try to do in this code, please let me know!

Answer 1

Using your code, I think this will give the effect you want. Use flatten() to remove the list hierarchy, and walk() to iterate and print the results for their side-effects only.

library(purrr)

#Return the plots and tables
x <- TypeLoop(PlotNTable)

walk(flatten(x), print)

Answer 2

Using c instead of list should ease the pain a bit.

TypeLoop <- function(fun) {
  all <- fun(c("setosa","versicolor","virginica"),"All")
  set <- fun("setosa","Setosa")
  vers <- fun("versicolor","Versicolor")
  virg <- fun("virginica","Verginica")
  return( c(all, set, vers, virg) )
}

The rest I cannot say for sure cause I did not try it in a knitr environment. But the next thing I would try is something like this. It seems to work in the console. You should be able to figure it out from here.

output <- function(x) {
  if ("gg" %in% class(x)) { 
    print(x)
  } else if ("knitr_kable" %in% class(x)) {
    cat(unlist(x), "\n")
  }
}
sapply(TypeLoop(PlotNTable), output)

Is that what you were looking for?