pivot_longer tibble involving NaN, NA, numeric(0)

Question

I'm trying to convert a tibble to a longer format.

library(tidyverse)

tbl <- tibble(
  a_X_values = list(numeric(0), c(3, NaN), c(4, 5, 6, NaN, NaN)),
  a_Y_values = list(5, NaN, numeric(0)),
  b_X_values = c(7, 8, 9),
  b_Y_values = c(NA, NaN, 1)
)

tbl_longer <- tbl %>% 
  unnest(where(is.list)) %>% 
  pivot_longer(cols = everything(), 
               names_to = c("alpha", "beta", ".value"), names_sep = "_")

I am getting the following error:

Error: In row 3, can't recycle input of size 5 to size 0.
Run `rlang::last_error()` to see where the error occurred.

How do I fix the error?

1. It should involve unnesting selected columns which are lists unnest(where(is.list))

2. A possible solution should also include converting all numeric(0) and NaN to NA.

Answer 1

Problem seems to be with the empty numeric values that are contained in lists, if they weren't listed your data.frame wouldn't work

If we apply a function over the list, searching for empty numeric values, replacing them with NA we can then use your code to pivot longer

library(dplyr)

tbl %>% 
  rowwise() %>% 
  mutate(across(where(is.list), \(x) ifelse(length(x) == 0, list(NA), list(x)))) %>% 
  unnest(where(is.list)) %>% 
  pivot_longer(cols = everything(), 
               names_to = c("alpha", "beta", ".value"), names_sep = "_") %>% 
  ungroup()

  alpha beta  values
   <chr> <chr>  <dbl>
 1 a     X         NA
 2 a     Y          5
 3 b     X          7
 4 b     Y         NA
 5 a     X          3
 6 a     Y        NaN
 7 b     X          8
 8 b     Y        NaN
 9 a     X        NaN
10 a     Y        NaN