incorrect output when using pivot_longer with values_drop_na = TRUE

[wjchulme]

There appears to be an issue with the values_drop_na = TRUE option when using pivot_longer with dtplyr. The names_to columns end up as odd character-rendered list-cols things, rather than simple vectors. See example:

library('dtplyr')
library('dplyr')
library('tidyr')

example.data <- data.frame(
  pid = c(1,2,3,1,2,3),
  valueA_1 = runif(6),
  valueA_2 = NA_real_,
  valueB_1 = runif(6),
  valueA_2 = NA_real_,
  type_1 = c("A","B", "C", "C", "B", "B"),
  type_2 = NA_character_
)

example.data %>%
  lazy_dt() %>%
  pivot_longer(
    cols = -pid,
    names_to = c(".value", "index"),
    names_pattern = "^(.*)_(\\d+)",
    values_drop_na = TRUE,
  ) %>% 
  as_tibble()

Which produces:

# A tibble: 6 × 5
    pid index                                                                                   valueA valueB type 
  <dbl> <chr>                                                                                    <dbl>  <dbl> <chr>
1     1 "c(\"1\", \"1\", \"1\", \"1\", \"1\", \"1\", \"2\", \"2\", \"2\", \"2\", \"2\", \"2\")"  0.266 0.945  A    
2     2 "c(\"1\", \"1\", \"1\", \"1\", \"1\", \"1\", \"2\", \"2\", \"2\", \"2\", \"2\", \"2\")"  0.372 0.661  B    
3     3 "c(\"1\", \"1\", \"1\", \"1\", \"1\", \"1\", \"2\", \"2\", \"2\", \"2\", \"2\", \"2\")"  0.573 0.629  C    
4     1 "c(\"1\", \"1\", \"1\", \"1\", \"1\", \"1\", \"2\", \"2\", \"2\", \"2\", \"2\", \"2\")"  0.908 0.0618 C    
5     2 "c(\"1\", \"1\", \"1\", \"1\", \"1\", \"1\", \"2\", \"2\", \"2\", \"2\", \"2\", \"2\")"  0.202 0.206  B    
6     3 "c(\"1\", \"1\", \"1\", \"1\", \"1\", \"1\", \"2\", \"2\", \"2\", \"2\", \"2\", \"2\")"  0.898 0.177  B    
Warning message:
In `[.data.table`(melt(`_DT27`, measure.vars = list(c("valueA_1",  :
  Coercing 'list' RHS to 'character' to match the type of column 2 named 'index'.

The correct output (which can be seen by removing the lazy_dt() line in the above code) should be:

# A tibble: 6 × 5
    pid index valueA valueB type 
  <dbl> <chr>  <dbl>  <dbl> <chr>
1     1 1      0.266 0.945  A    
2     2 1      0.372 0.661  B    
3     3 1      0.573 0.629  C    
4     1 1      0.908 0.0618 C    
5     2 1      0.202 0.206  B    
6     3 1      0.898 0.177  B