Is there a way to call variables from another file in R without having them appear in the workspace?

Question

I have a list of HEX colours that I want to use for my graphs/tables etc in R.

I have written a piece of code that calls these values at the start of the script.

col1 <- '#00573F'
col2 <- '#40816F'
col3 <- '#804B9F' 
col4 <- '#C0D5D0'
col5 <- '#A29161'

I then call these values when plotting throughout, for example:

x <- seq(-pi, pi, 0.1)
plot(x, sin(x),
     main="The Sine Function",
     ylab="sin(x)",
     type="l",
     col=col1)

This works perfectly.

However, I was wondering if there is a way to store these colour variables within R as a standard set of variables that I don't have to call every time I start a new script?

Also, it would be great if they didn't show up in the Environment as values purely because there are so many of these colours and I have a hard time keeping track of all the other values in there.

Answer 1

Many have adopted packages as the default way to write R code, to enable organising things like this.

You can get away with a barebone version, which I'll describe here.

You need a R/ folder; dir.create("R"). This directory should not contain scripts, but rather standalone functions, etc. that you have no problem sourcing whenever appropriate.

Inside of this you could make a custom_colors function; file.edit("R/custom_colors.R") (this will open a file in RStudio). Add:

custom_colors <- function(color_id) {
  c(
    col1 = '#00573F',
    col2 = '#40816F',
    col3 = '#804B9F',
    col4 = '#C0D5D0',
    col5 = '#A29161'
  )[color_id]
}

Then wherever you need it, you may write source("R/custom_colors.R") to have that single function enter your environment.

Thus you may call custom_colors(1) instead of col1.

Answer 2

A handful of options to consider

Develop an internal package for your color constants

I won't so far as to write the package, but packages may contain any R object (not just functions and data). You could develop an internal package to hold your color constants. If your package is names myInternals, you can then call

x <- seq(-pi, pi, 0.1)
plot(x, sin(x),
     main="The Sine Function",
     ylab="sin(x)",
     type="l",
     col= myInternals::col1)

If you have multiple people that need access to your constants, this is the path I would take. It's a bit more overhead work, but separates the constants into a separate environment that is relatively easy to access.

Truth be told, I have an internal package where I work now that uses @Mossa's strategy.

Use 'hidden objects'

If you precede an object with a ., it won't show up in the list of items in the environment (assuming you're using the RStudio pane)

But run the following:

.col1 <- "#00573F"

# .col1 doesn't show up
ls()

# .col1 does show up
ls(all.names = TRUE)

x <- seq(-pi, pi, 0.1)
plot(x, sin(x),
     main="The Sine Function",
     ylab="sin(x)",
     type="l",
     col= .col1)

This is probably the easiest, in my opinion, and what I would do if no one else needed access to my constants.

Use a list

Much like @Mossa's answer, using a list will reduce the number of new objects shown in the environment to just 1.

col_list <- list(col1 = '#00573F'
                 col2 = '#40816F'
                 col3 = '#804B9F' 
                 col4 = '#C0D5D0'
                 col5 = '#A29161')

x <- seq(-pi, pi, 0.1)
plot(x, sin(x),
     main="The Sine Function",
     ylab="sin(x)",
     type="l",
     col=col_env$col1)

Use an environment

This also only adds one object to the environment, and stores the constants outside of the current environment. Using them isn't much different than using a list, however, so I'm not sure what exactly is gained.

col_env <- new.env()

assign("col1", "#00573F", col_env)

x <- seq(-pi, pi, 0.1)
plot(x, sin(x),
     main="The Sine Function",
     ylab="sin(x)",
     type="l",
     col=col_env$col1)

Answer 3

You can add them to your .Rprofile as list or a function (as Mossa suggests), that R will run at each startup.

See this post on how to find your .Rprofile.