Assign the name of the variable to a non NA in a data frame with multiple variables

user_12307748 0
Viewed 37

For example, my df is:

         >dfABy 
         A    B     C

         56   NA  NA
         NA   45  NA
         NA   77  NA 
         67   NA  12 
         NA   65  3

I want to achieve the following data frame

         >dfABy 
         A    B    C

         A    NA  NA
         NA   B   NA
         NA   B   NA 
         A    NA  C
         NA   B   C
sapply dataframe r for-loop function
3 Answers

Here is an option in base R. Convert the data into a logical matrix with TRUE for non-NA and FALSE for NA. Replicate the column names based on the colum index ('nm1'). Assign the elements in the data based on the index 'i1' with the corresponding column names

i1 <- !is.na(dfABy)
nm1 <- names(dfABy)[col(dfABy)]
dfABy[i1] <- nm1[i1]

-output

dfABy
#     A    B    C
#1    A <NA> <NA>
#2 <NA>    B <NA>
#3 <NA>    B <NA>
#4    A <NA>    C
#5 <NA>    B    C

Or in a single line

dfABy[] <- names(dfABy)[col(dfABy)][(NA^is.na(dfABy)) * col(dfABy)]

Or using tidyverse

library(dplyr)
dfABy %>%
    mutate(across(everything(), ~ replace(., !is.na(.), cur_column())))
#     A    B    C
#1    A <NA> <NA>
#2 <NA>    B <NA>
#3 <NA>    B <NA>
#4    A <NA>    C
#5 <NA>    B    C

data

dfABy <- structure(list(A = c(56L, NA, NA, 67L, NA), B = c(NA, 45L, 77L, 
NA, 65L), C = c(NA, NA, NA, 12L, 3L)), class = "data.frame", row.names = c(NA, 
-5L))
user_37322710

The other alternative is to use the purrr package:

library(purrr)
 
df <- data.frame(A = c(56, NA, NA, 67, NA), B = c(NA, 45, 77, NA, 65), C = c(NA, NA, NA,12, 3))
 
imap_dfc(df, ~ifelse(is.na(.x), NA, .y))
# A tibble: 5 x 3
  A     B     C    
  <chr> <chr> <chr>
1 A     NA    NA   
2 NA    B     NA   
3 NA    B     NA   
4 A     NA    C    
5 NA    B     C
user_120510810

Base R option with Map :

dfABy[] <- Map(function(x, y) ifelse(is.na(x), NA, y), dfABy, names(dfABy))
dfABy

#     A    B    C
#1    A <NA> <NA>
#2 <NA>    B <NA>
#3 <NA>    B <NA>
#4    A <NA>    C
#5 <NA>    B    C
user_39629140
Related