Convert factor levels into columns and columns into factor levels

I have a dataset concerning a question for ranking the reasons for doing a PhD (for example).

df <- data.frame(
  id = c(1:4),
  rank1 = c("Salary", "Interest", "Career", "Title"),
  rank2 = c("Title", "Career", "Salary", NA),
  rank3 = c("Interest", "Title", NA, NA),
  rank4 = c("Career", NA, NA, NA))

> df
  id    rank1  rank2    rank3  rank4
1  1   Salary  Title Interest Career
2  2 Interest Career    Title   <NA>
3  3   Career Salary     <NA>   <NA>
4  4    Title   <NA>     <NA>   <NA>

The Person with id 1 rated "Salary" as most important reason, then "Title", etc...

However, I am trying to convert the factor levels of the variable to a column and the columns as the factor level of the variable in order to get this:

  id Salary Title Interest Career
1  1  rank1 rank2    rank3  rank4
2  2   <NA> rank3    rank1  rank2
3  3  rank2  <NA>     <NA>  rank1
4  4   <NA> rank1     <NA>   <NA>

Is there a way to do this in R? I have tried spread() from tidyr, but this is not what I am aiming for.
Any help is appreciated!
Thank you!

I believe @Chamkrai has the answer you're looking for (currently deleted) but I was thinking about what to do with the NA's. In this example you can replace the NA for id2 with "Salary", as this is the only one missing value for that id. You could also fill in the other NA's by sampling from the 'missing' values. I haven't been able to work out a succinct approach, but there's a small chance that this will help with your actual use-case:

library(tidyverse)

df <- data.frame(
  id = c(1:4),
  rank1 = c("Salary", "Interest", "Career", "Title"),
  rank2 = c("Title", "Career", "Salary", NA),
  rank3 = c("Interest", "Title", NA, NA),
  rank4 = c("Career", NA, NA, NA))
df
#>   id    rank1  rank2    rank3  rank4
#> 1  1   Salary  Title Interest Career
#> 2  2 Interest Career    Title   <NA>
#> 3  3   Career Salary     <NA>   <NA>
#> 4  4    Title   <NA>     <NA>   <NA>

unique_values <- df %>%
  select(-id) %>%
  pivot_longer(everything()) %>%
  na.omit() %>%
  distinct(value) %>%
  pull(value)
unique_values
#> [1] "Salary"   "Title"    "Interest" "Career"

df %>%
  t %>%
  as.data.frame %>%
  mutate(across(everything(), 
                ~ifelse(is.na(.x) & row_number() > 1,
                        unique_values[!(unique_values %in% .x)],
                        .x))) %>%
  t %>%
  as.data.frame %>%
  pivot_longer(-id) %>%
  pivot_wider(names_from = value,
              values_from = name)
#> # A tibble: 4 × 5
#>   id    Salary Title Interest Career
#>   <chr> <chr>  <chr> <chr>    <chr> 
#> 1 1     rank1  rank2 rank3    rank4 
#> 2 2     rank4  rank3 rank1    rank2 
#> 3 3     rank2  rank4 rank3    rank1 
#> 4 4     rank3  rank1 rank4    rank2

^{Created on 2023-06-15 with reprex v2.0.2}

library(tidyverse)
(df <- data.frame(
  id = c(1:4),
  rank1 = c("Salary", "Interest", "Career", "Title"),
  rank2 = c("Title", "Career", "Salary", NA),
  rank3 = c("Interest", "Title", NA, NA),
  rank4 = c("Career", NA, NA, NA)))

(df_long <- pivot_longer(df,
                        cols=-id) |> na.omit())

(df_rewide <- pivot_wider(data = df_long,
                          id_cols = "id",
                          names_from = "value",
                          values_from = "name"))