R 4.5: importing only selected objects & functions from R packages

On Friday, R version 4.5 was released. If you are curious about what’s new and what’s changed, you might enjoy reading Russ Hyde’s blog post or R’s official list of changes. For me, the introduction of the use() function is the most interesting aspect of this update.

Importing only selected objects & functions from a package

Until now, we had two choices of attaching a package to our R sessions:

Using library("dplyr") or require("poorman"), adding all functions, classes and objects that are exported by the package to our R session.
Using the triple-colon notation, e.g. dplyr::filter, and ensuring that the dependency is captured elsewhere (e.g. in the DESCRIPTION file of an R package)

The former can lead to unwanted surprises, e.g. when two packages export functions with the same name. (For example, dplyr's or poorman's filter functions mask base R’s filter function.)

Other programming languages, including python, allow users to attach / import only those functions or objects they want into the current workspace. For example, in python you might see: from math import pi

💡 Starting with R 4.5, you can now do the same in R - with the base::use() function. For example, the following code will only attach the select and mutate functions to your R session (but not dplyr’s filter function - or anything else):

use("poorman", c("filter", "mutate"))

No more wondering which filter function you are actually calling!

Buyer beware

But, as Konrad Rudolph points out, the new base::use() implementation is not perfect. Under the hood, it still uses library() - and as a consequence, you can only import objects from the same R package once per R session.

For example, running the following code snippet in a single R session leads to an error, because the second use() call is ineffective (but doesn’t tell you!):

use("poorman", c("filter"))  
filter(mtcars, cyl \> 3\)  \# works\!

use("poorman", c("select"))  
select(mtcars, cyl)  \# fails  
Error in select(mtcar, cyl) : could not find function "select"

So: buyer beware!

If you are interested in a different implementation of the use() approach, check out Konrad’s own box R package. Not only does it come with its own box::use() function, but it also allows you turn your own R code into reusable modules - without having to compose an entire R package every time!

This work is licensed under a Creative Commons Attribution 4.0 International License.

--- title: "R 4.5: importing only selected objects & functions from R packages" author: "Thomas Sandmann" date: "2025-04-12" execute: echo: true warning: false message: false draft: false freeze: true categories: [R] editor: markdown: wrap: 72 format: html: anchor-sections: true toc: true toc-depth: 4 code-tools: source: true toggle: false caption: none editor_options: chunk_output_type: console --- On Friday, [R version 4.5](https://cran.r-project.org/) was released. If you are curious about what's new and what's changed, you might enjoy reading [Russ Hyde's blog post](https://www.jumpingrivers.com/blog/whats-new-r45/) or R's official [list of changes](https://cran.r-project.org/bin/windows/base/NEWS.R-4.5.0rc.html). For me, the introduction of the `use()` function is the most interesting aspect of this update. ### Importing only selected objects & functions from a package Until now, we had two choices of attaching a package to our R sessions: * Using `library("dplyr")` or `require("poorman")`, adding all functions, classes and objects that are exported by the package to our R session. * Using the triple-colon notation, e.g. `dplyr::filter`, and ensuring that the dependency is captured elsewhere (e.g. in the `DESCRIPTION` file of an R package) The former can lead to unwanted surprises, e.g. when two packages export functions with the same name. (For example, `dplyr's` or `poorman's` `filter` functions mask base R's `filter` function.) Other programming languages, including python, allow users to attach / import only those functions or objects they want into the current workspace. For example, in python you might see: from math import pi 💡 Starting with R 4.5, you can now do the same in R \- with the `base::use()` function. For example, the following code will only attach the select and mutate functions to your R session (but not dplyr's filter function \- or anything else): ```r use("poorman", c("filter", "mutate")) ``` No more wondering which filter function you are actually calling\! ### Buyer beware But, as Konrad Rudolph [points out](https://mastodon.social/@klmr/114261539783693441), the new `base::use()` implementation is not perfect. Under the hood, it still uses `library()` \- and as a consequence, you can only import objects from the same R package *once* per R session. For example, running the following code snippet in a single R session leads to an error, because the second use() call is ineffective (but doesn't tell you\!): ```r use("poorman", c("filter")) filter(mtcars, cyl \> 3\) \# works\! use("poorman", c("select")) select(mtcars, cyl) \# fails Error in select(mtcar, cyl) : could not find function "select" ``` So: buyer beware\! If you are interested in a different implementation of the `use()` approach, check out [Konrad's](https://mastodon.social/@klmr) own [box R package](https://cran.r-project.org/package=box). Not only does it come with its own `box::use()` function, but it also allows you turn your own R code into reusable modules \- without having to compose an entire R package every time!