One of the packages I use all the time now in my interactive R sessions is pacman. It’s more than just a nicer way to load packages, it’s a great way to explore packages, functions, package libraries, versioning, etc.

Grab it off cran with install.packages("pacman").

Out of the box, the function I use every day in all my scripts or at the console is pacman::p_load(); it’s a great substitute for library() or require(). Yihui has already written about his preferences/understanding on the library vs. require.

I prefer pacman::p_load() for different reasons than he describes. Instead of having to write 5 lines of code to load for 5 common data munging packages, like so:

library(dplyr)
library(tidyr)
library(plyr) 
library(magrittr)
library(ggplot2)

You can write one line

pacman::p_load(dplyr,tidyr,plyr,ggplot2, magrittr)

If you want to unload a package, say to update it, the function pacman::p_unload() is pretty useful. Even better pacman::p_load() will install and load new packages in the same command. And for github or bitbucket accounts, use pacman::p_load_gh().

Yes, I know that if brevity in syntax for package loading were my main motivation, I could just do import pandas for python or library(data.table) in R and get the equivalent functionality in one package load; but brevity of package loading isn’t my main concern in language or package choosing.

Something else that’s really nice is a way to print all the functions available in a package. Try it with pacman::p_funs(magrittr). You may find the print out overwhelming for a large packages like pacman::p_funs(stats); but I still like it in cases when IntelliSense auto-complete doesn’t quite get me to the function I’m looking for or is cramped for reading.

Also, whenever you’re doing a update to R or you just want to quickly back up all your packages, I find pacman::p_lib() support helpful. For example,

writeLines(pacman::p_lib(), "~/Desktop/list_of_R_packages.csv")

Lastly, to find out what the path to your R-library, the function pacman::p_path() is a real help. Check out the pacman github page for some more details.

In the next blog post, we’ll have a bit of fun with

Tweet