Create a nested list from a dataframe using multiple columns

Question

I have a dataframe that I would like to convert to a nested list, by row. However, I need to group several columns into separate matrices in each nested element. All of the solutions I've seen put each column into a unique element. I want to create separate matrices using the column values.

For example, if I have a dataframe as below:

df = data.frame(a=c("foo", "bar", "baz"), b=1:3, c=3:1, d=4:6, e=6:4)

I'd like to convert it to something like this:

dflist <- list(foo=list(as.matrix(c(1,3)),
                        as.matrix(c(4,6))),
               bar=list(as.matrix(c(2,2)),
                        as.matrix(c(5,5))),
               baz=list(as.matrix(c(3,1)),
                        as.matrix(c(6,4))))

jpsmith · Accepted Answer · 2024-08-26 21:52:35Z

2

In base R, you can split and then create your list for each using lapply:

ll <- lapply(split(df, df$a), \(x) list(matrix(x[2:3]), matrix(x[4:5])))

#If order matters (above is alphabetical (bar, bar, foo):
ll_ordered <- ll[match(df$a, names(ll))]

Or all in one line, thanks to the great advice by @Friede:

lapply(split(df, factor(f <- df$a, f)), \(x) list(matrix(x[2:3]), matrix(x[4:5])))

Both approaches give the out output:

# $foo
# $foo[[1]]
# [,1]
# [1,] 1   
# [2,] 3   
# 
# $foo[[2]]
# [,1]
# [1,] 4   
# [2,] 6   
# 
# 
# $bar
# $bar[[1]]
# [,1]
# [1,] 2   
# [2,] 2   
# 
# $bar[[2]]
# [,1]
# [1,] 5   
# [2,] 5   
# 
# 
# $baz
# $baz[[1]]
# [,1]
# [1,] 3   
# [2,] 1   
# 
# $baz[[2]]
# [,1]
# [1,] 6   
# [2,] 4

edited Aug 26, 2024 at 21:52

answered Aug 26, 2024 at 18:07

jpsmith

18.5k6 gold badges25 silver badges48 bronze badges

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

Comments

Axeman · Accepted Answer · 2024-08-26 17:50:17Z

1

Using purrr:

mylist <- pmap(df, \(b, c, d, e, ...) {
  list(as.matrix(c(b, c)), as.matrix(c(d, e)))
}) |> setNames(df$a)

answered Aug 26, 2024 at 17:50

Axeman

35.7k8 gold badges87 silver badges101 bronze badges

Comments

G. Grothendieck · Accepted Answer · 2024-08-26 18:32:54Z

0

For each row create the list of matrices in a new column, data, and then extract that column. Note that in the question the input is integer but dflist contains doubles. We assume that was not intended and use dflist as defined in the Note at the end instead.

library(dplyr)

result <- df %>%
  rowwise %>%
  mutate(data = list(cbind(c(b, c)), cbind(c(d, e))) %>%
    list %>%
    setNames(a)) %>%
  pull(data)

identical(result, dflist)
## [1] TRUE

Note

dflist <- list(foo=list(as.matrix(c(1L,3L)),
                        as.matrix(c(4L,6L))),
               bar=list(as.matrix(c(2L,2L)),
                        as.matrix(c(5L,5L))),
               baz=list(as.matrix(c(3L,1L)),
                        as.matrix(c(6L,4L))))

edited Aug 26, 2024 at 18:32

answered Aug 26, 2024 at 18:22

G. Grothendieck

273k18 gold badges220 silver badges365 bronze badges

Collectives™ on Stack Overflow

Create a nested list from a dataframe using multiple columns

3 Answers 3

Comments

Comments

Note

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

Note

Comments

Your Answer

Sign up or log in

Post as a guest

Related