Replace value by column name for many columns using R and dplyr [duplicate]

Question

I am transforming the results of a survey, including multiple selection responses. The original data looks like this:

df <- data_frame(
  id = c("a", "b", "c"),
  tired = c(T, F, T),
  lonely = c(F, F, T),
  excited = c(F, T, T)
)

df
# A tibble: 3 x 4
  id    tired lonely excited
  <chr> <lgl> <lgl>  <lgl>  
1 a     TRUE  FALSE  FALSE  
2 b     FALSE FALSE  TRUE   
3 c     TRUE  TRUE   TRUE   

I would like to create a new column "feelings" that contains comma separated values of the feelings expressed by a respondent:

  id    feelings              
  <chr> <chr>                 
1 a     tired, excited        
2 b     excited               
3 c     tired, lonely, excited

An intermediate step would be to replace TRUE values with the respective name of the column in order to yield:

  id    tired  lonely   excited
  <chr> <lgl>  <lgl>    <lgl>  
1 a     tired          excited  
2 b                    excited   
3 c     tired  lonely  excited   

For an individual column this is straightforward. However, unlike the example, there are a lot of columns in my data frame (10+, with usually no more than one or two TRUE values), and therefore I would like to automate this for a number of columns. One solution would probably be to loop over the columns and use base subsetting and replacement, but is there also an elegant dplyr/tidy way to do this?

Thanks for your help!

Answer 1

An option is to use tidyr::gather and then summarise using dplyr :

library(dplyr)
library(tidyr)
df %>% gather(feelings, value, -id) %>%  #Change to long format
  filter(value) %>%  #Filter for value which are TRUE
  group_by(id) %>%
  summarise(feelings= paste0(feelings,collapse=","))  

# id    feelings
# <chr> <chr>               
# 1 a     tired               
# 2 b     excited             
# 3 c     tired,lonely,excited