`bind_rows()` shifts named list identifiers when multiple consecutive NULL elements are present

[koncina]

When row-binding a named list containing multiple consecutive NULL elements the resulting data frame identifiers (using the .id argument) are shifted:

library(tidyverse)

list(a = tibble(expected_id = "a"),
     b = NULL,
     c = NULL,
     d = tibble(expected_id = "d"),
     c = NULL,
     e = tibble(expected_id = "e")) %>% 
  bind_rows(.id = "id")
#> # A tibble: 3 x 2
#>   id    expected_id
#>   <chr> <chr>      
#> 1 a     a          
#> 2 c     d          
#> 3 d     e

^{Created on 2019-03-20 by the reprex package (v0.2.1)}

Single NULL elements are handled accurately:

library(tidyverse)

list(a = tibble(expected_id = "a"),
     b = NULL,
     c = tibble(expected_id = "c"),
     d = NULL,
     e = tibble(expected_id = "e")) %>% 
  bind_rows(.id = "id")
#> # A tibble: 3 x 2
#>   id    expected_id
#>   <chr> <chr>      
#> 1 a     a          
#> 2 c     c          
#> 3 e     e

^{Created on 2019-03-20 by the reprex package (v0.2.1)}

[werkstattcodes]

just noted that using it with purrr leads to desired result

> list(a = tibble(expected_id = "a"),
+      b = NULL,
+      c = NULL,
+      d = tibble(expected_id = "d"),
+      f = NULL,
+      e = tibble(expected_id = "e")) %>% 
+   map_dfr(., bind_rows, .id="id")
# A tibble: 3 x 2
  id    expected_id
  <chr> <chr>      
1 a     a          
2 d     d          
3 e     e   

[koncina]

But only because bind_rows() coerces each single NULL element to an empty tibble (just like as_tibble()). map_dfr() per se will also fail to handle indices of multiple NULL elements:

library(tidyverse)

bind_rows(NULL) %>% 
  class()
#> [1] "tbl_df"     "tbl"        "data.frame"

my_list <- list(a = tibble(expected_id = "a"),
     b = NULL,
     c = NULL,
     d = tibble(expected_id = "d"),
     c = NULL,
     e = tibble(expected_id = "e"))

map_dfr(my_list, as_tibble, .id = "id")
#> # A tibble: 3 x 2
#>   id    expected_id
#>   <chr> <chr>      
#> 1 a     a          
#> 2 d     d          
#> 3 e     e

map_dfr(my_list, ~.x, .id = "id")
#> # A tibble: 3 x 2
#>   id    expected_id
#>   <chr> <chr>      
#> 1 a     a          
#> 2 c     d          
#> 3 d     e

^{Created on 2019-03-21 by the reprex package (v0.2.1)}

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/