R: Merge lists containing vectors and scalars

Question

I have N lists that have some identical column names. Here is a MWE with two list:

ls <- list()
ls[[1]] <- list("a"=1:2,
                "b"=20,
                "c"=numeric(0))
names(ls[[1]]$a) <- c("a1", "a2")

ls[[2]] <- list("a"=3:4,
                "b"=30,
                "c"=1:4,
                "d"="f")
names(ls[[2]]$a) <- c("a1", "a2")

Is it possible merge these into a resulting list lsRes, where lsRes has the following properties:

1. lsRes$a contains two elements, where the first is the named vector c(1,2) (with names c(a1, a2)) and the second a named vector c(3,4) (with names (c(a1.a2)))

2. lsRes$b contains two elements, where the first is 20 and the second is 30

3. lsRes$c contains two elements, where the first is numeric(0) and the second is 1:4

4. lsRes$d contains two elements, where the first is NA and the second is "f"

I looked at this and this, but they describe different cases

Answer 1

Assuming that we need to have the output also as a list, we create the common names and then assign those doesn't any of the common names to NA

nm1 <- unique(unlist(sapply(ls, names)))
lsRes <- lapply(ls, function(x) {x[setdiff(nm1, names(x))] <- NA; x})
lengths(lsRes)
#[1] 4 4

If we need to have a list of 4 elements, then use transpose

library(purrr)
lsRes %>%
      transpose