too_many

df |> 
  separate_wider_delim(
    x, "-", names = c("x", "y"),
    too_few = "align_start",
    too_many = "merge"
  )

## # A tibble: 3 × 2
##   x     y    
##   <chr> <chr>
## 1 1     1-1  
## 2 1     3    
## 3 1     <NA>