README.md

In philstraforelli/forcatsmisc: Additional functions for forcats

forcatsmisc

Factor variables have long been a pain in handling in R. While the forcats package has alleviated many of the difficulties associated in handling factor variables, I have found from time to time other needs in wrangling those variables which are unworthy of belonging to the forcats package. They will therefore live here instead.

Installation

You can install forcatsmisc from github with:

# install.packages("remotes")
remotes::install_github("philstraforelli/forcatsmisc")

Examples

fct_recode_if()

It’s sometimes handy to recode a factor level on a predicate, particularly when used in a shiny reactive environment.

df <- data.frame(x = factor(c("apple", "bear", "banana", "dear")),
                 y = factor(c("fruit", "animal", "fruit", "animal")))

dplyr::mutate(df, x = forcatsmisc::fct_recode_if(x, y == "animal", "not a fruit"))
#>             x      y
#> 1       apple  fruit
#> 2 not a fruit animal
#> 3      banana  fruit
#> 4 not a fruit animal

fct_to_num()

One of the most upvoted R questions on StackOverflow is a way to handle situations when a numeric variable was erroneously convered to a factor variable, and you want it converted back to numeric - A frustrating situation that occurs surprisingly often. It’s possible that the question’s popularity may have led to the decision that it wasn’t worth formalizing the solution into its own function, I decided otherwise and formalized it into this function.

var <- factor(c("1", "2", "3"))

forcatsmisc::fct_to_num(var)
#> [1] 1 2 3

fct_wrap()

One annoyance I often run into is that I want to use a factor variable as a categorical axis in a plot, but one or more of the factor levels is very long. I then make the mistake of using str_wrap(), which converts my factor variable back to a string variable, thus losing the order of levels I have set earlier! Grrrr… Well, in such situations, use fct_wrap() instead.

var <- factor(c("a very, very, very, very long factor level", "an even longer, unbelievably longer, factor level"))

forcatsmisc::fct_wrap(var, width = 10)
#> [1] a very,\nvery,\nvery,\nvery long\nfactor\nlevel       
#> [2] an even\nlonger,\nunbelievably\nlonger,\nfactor\nlevel
#> 2 Levels: a very,\nvery,\nvery,\nvery long\nfactor\nlevel ...

fct_trim()

Another annoyance is the inclusion of whitespace in factor levels. fct_trim() is the factor level version of str_trim() to answer to this issue.

var <- factor(c("  Level with trailing and leading white space\t", "\n\nLevel with trailing and leading white space\n\n"))

forcatsmisc::fct_trim(var)
#> [1] Level with trailing and leading white space
#> [2] Level with trailing and leading white space
#> Levels: Level with trailing and leading white space

lvl_range()

One thing I’ve been doing more and more often in a shiny app is to place the scales of a factor variable using shinyWidgets::sliderTextInput() and let the user select a range they want displayed. The lvl_range() was created to make it easier to handle the output of the UI function on the backend (such as when filtering the data to include only the range of factor levels selected).

var <- factor(c("Very likely", "Somewhat likely", "Somewhat unlikely", "Very unlikely"), 
              levels = c("Very likely", "Somewhat likely", "Somewhat unlikely", "Very unlikely"))

forcatsmisc::lvl_range(var, c("Very likely", "Somewhat unlikely"))
#> [1] "Very likely"       "Somewhat likely"   "Somewhat unlikely"

lvl_number()

I’ve lately found a need to return the position of a particular level in a factor variable. For instance, in levels(iris$Species), the “versicolor” level is in the 2nd position. This function makes it easier to return this.

levels(iris$Species)
#> [1] "setosa"     "versicolor" "virginica"
forcatsmisc::lvl_number(iris$Species, "versicolor")
#> [1] 2

fct_label_counts()

This function easily adds counts of factors in their levels. This is useful for displaying in a chart how many cases there are in each level.

var <- forcatsmisc::fct_label_counts(iris$Species)
levels(var)
#> [1] "setosa (n=50)"     "versicolor (n=50)" "virginica (n=50)"

philstraforelli/forcatsmisc documentation built on July 4, 2023, 9:52 a.m.