Force lapply/sapply to index output list with inputs

Question

As in the title, how to force lapply or sapply to use input vector values to index the output. I have two cases; let me begin with the simpler one. The following code

lapply(c('a', 'b'), function(idx){
    rnorm(1)
})

gives me the output

[[1]]
[1] -1.359386

[[2]]
[1] -0.3428958

but I would like to receive:

$a
[1] -1.359386

$b
[1] -0.3428958

In the final aim, my code looks like follows:

m = sapply(unique(isotopes[, 1]), function(el){
    sapply(isotopes[isotopes$element == el, 2], function(mnr){
        isotopes[isotopes$element == el & isotopes$mass_nr == mnr, 3]
    })
})

and it gives me outputs like

$Pu
[1] 238.0496 239.0522 240.0538 241.0569 242.0587 244.0642

while I would like to obtain list of lists, to be able to use m['Pu'][240] or m['Pu']['240'] etc. (for 'Pu' isotopes[isotopes$element == el, 2] is a vector of 238 239 240 241 242 244)

PS. Of course, I can use

m = function(el, mnr){ isotopes[isotopes$element == el & isotopes$mass_nr == mnr, 3] }

but I'm interested if the above problem possesses a neat solution :)

Answer 1

You could try

lapply(c(a = "a", b = "b"), function(idx) {
    rnorm(1)
})

which should give the desired output

$a
[1] 1.556433

$b
[1] -0.2282002

Answer 2

If you want to split the data set twice, use split twice, it returns a named list.

In the code below I have used a column name third_col, not the column number.

set.seed(2023)
n <- 20L
isotopes <- data.frame(
  element = sample(letters[1:2], n, TRUE),
  mass_nr = sample(3, n, TRUE),
  third_col = rnorm(n)
)


sp <- split(isotopes, isotopes$element) 
lapply(sp, \(x) split(x$third_col, x$mass_nr))
#> $a
#> $a$`1`
#> [1]  0.8602705 -0.1299832  0.1092238  0.1615685 -1.5131984  1.1431521
#> 
#> $a$`2`
#> [1] -0.4709085 -0.4389865 -1.1927195 -1.1326977
#> 
#> $a$`3`
#> [1]  0.8468178 -1.3264019  0.9712922
#> 
#> 
#> $b
#> $b$`1`
#> [1] 1.9957190 0.1108168
#> 
#> $b$`2`
#> [1]  0.7522100  0.9254184 -0.2917223
#> 
#> $b$`3`
#> [1] -0.2294296 -0.7137467

^{Created on 2023-09-20 with reprex v2.0.2}

Or with a pipe from the first split to a lapply/split. Then extract a couple of named list members.

m <- split(isotopes, isotopes$element) |> lapply(\(x) split(x$third_col, x$mass_nr))

m$a[['2']]
#> [1] -0.4709085 -0.4389865 -1.1927195 -1.1326977

m$b$`3`
#> [1] -0.2294296 -0.7137467

^{Created on 2023-09-20 with reprex v2.0.2}

Answer 3

Another way - use sapply() with the option simplify = FALSE to get a list as output:

sapply(c('a', 'b'), \(idx) rnorm(1), simplify = FALSE)

Output:

$a
[1] 0.4824588

$b
[1] 0.7582138