运行pivot_wider后处理NA

Question

I have a long dataframe that I want to make wide using pivot_wider :

library(tidyr)
example_data <- data.frame(
    name = c("bob", "bob", "dick", "dick", "harry", "harry"), 
    sport = c("baseball", "football", "hockey", "basketball", "football", "basketball")
)
pivot_wider(example_data, names_from = sport, values_from = sport)

This gives the expected result, but with lots of NA s

  name  baseball football hockey basketball
1 bob   baseball football NA     NA        
2 dick  NA       NA       hockey basketball
3 harry NA       football NA     basketball

I want to convert the sport names to TRUE (since the name of the sport is already indicated by the column name) and convert the NA s to FALSE , creating a dataframe like this:

   name baseball football hockey basketball
1   bob     TRUE     TRUE  FALSE      FALSE
2  dick    FALSE    FALSE   TRUE       TRUE
3 harry    FALSE     TRUE  FALSE       TRUE

I thought this code would do the trick, but it threw an error:

pivot_wider(
    example_data, 
    names_from = sport, 
    values_from = sport,
    values_fill = list(sport = FALSE),
    values_fn = list(sport = !is.na)
)
Error in !is.na : invalid argument type

The code below gets me the reverse of what I'm looking for, which I could then work to convert to the desired dataframe:

pivot_wider(
    example_data, 
    names_from = sport, 
    values_from = sport,
    values_fill = list(sport = TRUE),
    values_fn = list(sport = is.na)
)

Is there a way to get directly to the desired dataframe? And are there any tutorials for how to use the values_fn argument so I can figure out why values_fn = list(sport = !is.na) isn't working? Thanks.

Answer 1

One way would be to create a dummy column with TRUE values and then use pivot_wider .

library(dplyr)
library(tidyr)

example_data %>%
  mutate(val = TRUE) %>%
  pivot_wider(names_from = sport,values_from = val,values_fill = list(val = FALSE))


# A tibble: 3 x 5
#  name  baseball football hockey basketball
#  <fct> <lgl>    <lgl>    <lgl>  <lgl>     
#1 bob   TRUE     TRUE     FALSE  FALSE     
#2 dick  FALSE    FALSE    TRUE   TRUE      
#3 harry FALSE    TRUE     FALSE  TRUE