List names as variables

I often end up wanting to build a list with named items, and the names come in as variables.

For example, instead of the typical

list(a = 'first', b = 'second', d = c(3,4))

$a
[1] "first"

$b
[1] "second"

$d
[1] 3 4

I might have the names defined elsewhere. This is particularly common inside functions.

name1 <- 'a'
name2 <- 'b'
name3 <- 'd'

We can’t pass those in as usual- this uses name# as the name, not the value of the variable.

list(name1 = 'first', name2 = 'second', name3 = c(3,4))

$name1
[1] "first"

$name2
[1] "second"

$name3
[1] 3 4

Various solutions

We can use setNames

setNames(list('first', 'second',c(3,4)), c(name1, name2, name3))

$a
[1] "first"

$b
[1] "second"

$d
[1] 3 4

we can do it in two steps with names , which I think is what setNames wraps, and is just extra verbose and requires carrying data copies around.

barelist <- list('first', 'second',c(3,4))
names(barelist) <- c(name1, name2, name3)
barelist

$a
[1] "first"

$b
[1] "second"

$d
[1] 3 4

Can we unquote/eval?

I can almost never get !! or !!! to work. this doesn’t, as usual.

list(!!name1 = 'first', !!name2 = 'second', !!name3 = c(3,4))

Nor this

list(eval(name1) = 'first', eval(name2) = 'second', eval(name3) = c(3,4))

Nor this, despite the eval working

rlang::eval_bare(name1)

[1] "a"

list(rlang::eval_bare(name1) = 'first', rlang::eval_bare(name2) = 'second', rlang::eval_bare(name3) = c(3,4))

How about tibble::lst ? I often use it for lists of variables because it self-names them, so maybe it’s the answer here. Yep. That’s just cleaner.

tibble::lst(name1 = 'first', name2 = 'second', name3 = c(3,4))

$name1
[1] "first"

$name2
[1] "second"

$name3
[1] 3 4

And, that self-naming I was describing, which solves a different problem- having to write list(name = name, age = age).

name <- c('David', 'Susan')
age <- c(1,2)

tibble::lst(name, age)

$name
[1] "David" "Susan"

$age
[1] 1 2