I have a large df file from a survey platform. the columns contains one question(as a title) with all potential answer. I'm trying to separate each of this answer column into a new set of columns (with the name of answer as the the name of the new column).

after that i would like each new column to indicate if the value appears in the original column(with "1") to process the data more easily

     df<-data.frame("name"= c("John","mark","bell","elsa"),"what do you like to 
     eat"=c("apple","fries apple","peach","bread"))

original df

i'm using this code which works, but i'm sure there must be more efficient/easier way of doing it, as i have a over 50 column like this.

    df<-df %>%
  separate(what.do.you.like.to.eat, c("apple","fries","peach","bread",NA ), remove = F)
df[,3:6]<-""
{
  df[,3] =  with(df, ifelse(grepl("apple", df$what.do.you.like.to.eat,ignore.case = T), 
                                             paste('1', df[,3]), 
                                             paste("", df[,3])))
  df[,4] =  with(df, ifelse(grepl("fries", df$what.do.you.like.to.eat,ignore.case = T), 
                            paste('1', df[,4]), 
                            paste("", df[,4])))
  df[,5] =  with(df, ifelse(grepl("peach", df$what.do.you.like.to.eat,ignore.case = T), 
                            paste('1', df[,5]), 
                            paste("", df[,5])))
  df[,6] =  with(df, ifelse(grepl("bread", df$what.do.you.like.to.eat,ignore.case = T), 
                            paste('1', df[,6]), 
                            paste("", df[,6])))
}

desire output

CodePudding user response：

You could use purrr::map to apply over your vector of answers and for each, check their presence in the string.

library(tidyverse)

df <- data.frame(
  name = c("John", "mark", "bell", "elsa"),
  "what do you like to eat" = c("apple", "fries apple", "peach", "bread")
)

ans <- c("apple", "fries", "peach", "bread")

map_dfc(ans,~ transmute(df, !!sym(.x) := str_detect(what.do.you.like.to.eat, .x))) %>%
  bind_cols(df, .)
#>   name what.do.you.like.to.eat apple fries peach bread
#> 1 John                   apple  TRUE FALSE FALSE FALSE
#> 2 mark             fries apple  TRUE  TRUE FALSE FALSE
#> 3 bell                   peach FALSE FALSE  TRUE FALSE
#> 4 elsa                   bread FALSE FALSE FALSE  TRUE

CodePudding user response：

Ok i've done this tell me if this works for you :

my_df <- data.frame("name" = c("John","mark","bell","elsa"),
                "what do you like to eat" = c("apple","fries apple","peach","bread"),
                stringsAsFactors = FALSE)
my_var <- unique(sort(str_split(string = my_df$what.do.you.like.to.eat, pattern = " ", simplify = TRUE)))
my_pos <- which(my_var == "")
if (length(my_pos)) {
  my_var <- my_var[-my_pos]
}
my_col <- c(colnames(my_df), my_var)
my_miss <- setdiff(my_col, colnames(my_df))
my_df[my_miss] <- NA
my_f <- function(x, y) {
  my_var <- grep(pattern = colnames(my_df)[x], x = my_df[, y])
  if (length(my_var)) {
    my_df[my_var, x] <<- 1
  }
}
lapply(3:ncol(my_df), function(x) my_f(x, 2))

you can change this part to this :

my_df <- data.frame("name" = c("John","mark","bell","elsa"),
                "what do you like to eat" = c("i like apple","i love fries apple","i'm kind of peach","bread all the way"),
                stringsAsFactors = FALSE)
my_var <- unique(sort(str_split(string = 
my_df$what.do.you.like.to.eat, pattern = " ", simplify = TRUE)))
my_food <- c("apple", "fries", "bread", "peach")
my_var <- my_var[which(my_var %in% my_food)]
my_pos <- which(my_var == "")
if (length(my_pos)) {
  my_var <- my_var[-my_pos]
}