Convert NULLs to NAs in a list-CodePudding

Would you suggest a better (briefer or more legible) way of converting NULLs in a list to NAs; and from list to vector?

list(1, 2, 3, numeric(0), 5) %>%
  purrr::map_dbl(~ ifelse(length(.) == 0, NA_real_, .))
# [1]  1  2  3 NA  5

I would prefer not using ifelse and instead using if_else. Is there another way of doing it with purrr?

CodePudding user response：

If the length of each element of a list is 0 or 1 (e.g. lst.1), you could simply use

lst.1 %>% map_dbl(1, .default = NA)

# [1]  1  2  3 NA  5

A general way to deal with a list with different length in each element (e.g. lst.2) is

lst.2 %>%
  map_if(~ length(.) == 0, ~ NA) %>%
  flatten_dbl()

# [1]  1  2  3 NA  5

lst.1 <- list(1, 2, 3, numeric(0), 5)
lst.2 <- list(1:3, numeric(0), 5)

CodePudding user response：

If L is the list then any of these work:

replace(L, lengths(L) == 0, NA)

ifelse(lengths(L), L, NA)

No packages needed.