Using objects inside list as function arguments in lapply

Question

I am trying to learn different ways to use to the objects inside a list as the FUN arguments in lapply. Take this data:

A <- list(a = matrix(0, ncol = 3, nrow = 3), b = matrix(0, ncol = 3, nrow = 3))
B <- list(a = matrix(1, ncol = 1, nrow = 3), b = matrix(1, ncol = 1, nrow = 3))
D <- mapply(FUN="list", A, B, SIMPLIFY=F, USE.NAMES=F)
D <- lapply(D, `names<-`, c("first", "second"))
D
[[1]]
[[1]]$`first`
     [,1] [,2] [,3]
[1,]    0    0    0
[2,]    0    0    0
[3,]    0    0    0

[[1]]$second
     [,1]
[1,]    1
[2,]    1
[3,]    1


[[2]]
[[2]]$`first`
     [,1] [,2] [,3]
[1,]    0    0    0
[2,]    0    0    0
[3,]    0    0    0

[[2]]$second
     [,1]
[1,]    1
[2,]    1
[3,]    1

Desired result:

[[1]]
     [,1] [,2] [,3]
[1,]    0    0    0
[2,]    0    0    0
[3,]    0    0    0
[4,]    1    1    1

[[2]]
     [,1] [,2] [,3]
[1,]    0    0    0
[2,]    0    0    0
[3,]    0    0    0
[4,]    1    1    1

This is how I would normally do it:

lapply(D, function(x) rbind(x$first, as.numeric(x$second)))

Now I am wondering whether there is a way to avoid using function(x) and repeating all those xs. Something like:

lapply(D, "rbind", <args>)

How can I let rbind (or any other function) know that I am referring to objects within the frame of lapply?

Thank you,

K.

It's usually not possible if you nest several functions (e.g., rbind, as.numeric and $). — Roland
– Roland, Commented Aug 24, 2018 at 9:15
You can do something like lapply(D, do.call, what = cbind). The same would work with rbind if you didn't have a one-column matrix but a one-row matrix or vector. — Roland
– Roland, Commented Aug 24, 2018 at 9:18
@Roland using the same idea, tcbind <- function(...) t(cbind(...)); lapply(D, do.call, what = tcbind) — agstudy
– agstudy, Commented Aug 24, 2018 at 9:23
@agstudy Well, of course, there are practically no limitations if you define a function outside of lapply. But that is not much different from using an anonymous function within lapply. — Roland
– Roland, Commented Aug 24, 2018 at 9:26

Aurèle · Accepted Answer · 2018-08-24 15:24:54Z

2

To "avoid using function(x) and repeating all those xs", we could use with():

lapply(D, with, rbind(first, as.numeric(second)))

edited Aug 24, 2018 at 15:24

answered Aug 24, 2018 at 15:05

Aurèle

12.9k1 gold badge34 silver badges54 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

thothal · Accepted Answer · 2018-08-24 16:41:42Z

1

Update

As per the commments changed the code to keep only the right solution.

As some of the commentators said, the problem is that B would need to be transposed, to find an elegant solution. You should have a look to library(purrr) because with that the whole problem reduces to:

map2(A, B, ~ rbind(.x, t(.y)))
# $`a`
#      [,1] [,2] [,3]
# [1,]    0    0    0
# [2,]    0    0    0
# [3,]    0    0    0
# [4,]    1    1    1

# $b
#      [,1] [,2] [,3]
# [1,]    0    0    0
# [2,]    0    0    0
# [3,]    0    0    0
# [4,]    1    1    1

What map2 does is that it takes 2 lists and applies the function to each element of these lists. The ~ syntax is a shortcut for function(.)

edited Aug 24, 2018 at 16:41

answered Aug 24, 2018 at 11:21

thothal

20.7k4 gold badges44 silver badges88 bronze badges

2 Comments

Aurèle Over a year ago

In the same spirit, if we want to operate on D directly, we could: transpose(D) %>% pmap(~ rbind(.x, t(.y)))

Aurèle Over a year ago

Also, the two methods don't produce identical results when the matrix is not just 0s. The upper matrix is transposed with the first method

Collectives™ on Stack Overflow

Using objects inside list as function arguments in lapply

2 Answers 2

Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related