Spread colum into one-hot encoded pattern

Question

I have a data frame of the following pattern:

df <- tibble(ID = c(1, 1, 1, 2, 2), key = c("a", "b", "b", "c", "c"), value = c("k1", "k3", "k1", "k2", "k5"))
     ID key   value
  <dbl> <chr> <chr>
1     1 a     k1   
2     1 b     k3   
3     1 b     k1   
4     2 c     k2   
5     2 c     k5

What I need is for each ID, group together the rows where key is equal and then encode the value in a one-hot encoded manner accross all possible unique values of column value. That is I want s.th. like

     ID key      k1    k2    k3    k5
  <dbl> <chr> <dbl> <dbl> <dbl> <dbl>
1     1 a         1     0     0     0
2     1 b         1     0     1     0
3     2 c         0     1     0     1

I could provide a list of possible values like possible_values = c("k1", "k2", "k3", ...) if this helps

score 2 · Accepted Answer · answered May 23 '22 at 08:05

You can adjust those arguments in pivot_wider() to achive it.

library(tidyr)

pivot_wider(df,
            names_from = value, names_sort = TRUE,
            values_from = value, values_fn = length, values_fill = 0)

# A tibble: 3 × 6
     ID key      k1    k2    k3    k5
  <dbl> <chr> <int> <int> <int> <int>
1     1 a         1     0     0     0
2     1 b         1     0     1     0
3     2 c         0     1     0     1

Spread colum into one-hot encoded pattern

1 Answers1