Is there a function to update a dataframe?-CodePudding

I'm new to R and have tried looking for the correct way to write this. I apologize if this is a relatively rudimentary question!

all_loan_offers <- function(df){

df$decision <- c()
for(i in 1:nrow(df)) {

income <- df$income[i]
c_record <- df$c_record[i]
years <- df$years[i]
decision <- df$decision[i]

if (income > 80000){
  decision[i] <- "Admit"
} else if(income < 40000 & c_record == FALSE){
  decision[i] <- "Admit"
} else if(income < 40000 & c_record == TRUE){
  decision[i] <- "Admit"
} else if(income >= 40000 & income <= 80000 & years > 3){
  decision[i] <- "Admit"
} else {
  decision[i] <- "Reject"
}
return(df)

} }

applicants_info <- data.frame(income = c(40000, 80000, 25000, 70000),
                          c_record = c(F, T, T, F),
                          years = c(2, 10, 3, 6),
                          stringsAsFactors = F) 


all_loan_offers(applicants_info)

all_loan_offers then returns the applicants_info df, but my objective is to return an updated data frame with a 4th column to "Admit" or "Reject" based on my if-else ladder.

Many thanks - any help at all would be greatly appreciated! :)

CodePudding user response：

You probably don't need the local decision variable inside the for loop. When you set df$decision <- c() you are not doing anything, c() is NULL. You can set the initial value of all rows to "Reject" and then admit when conditions are met (dropping the last else { ... }).

all_loan_offers <- function(df) {
  
  df$decision <- "Reject"
  for(i in 1:nrow(df)) {
    
    income <- df$income[i]
    c_record <- df$c_record[i]
    years <- df$years[i]
    
    if (income > 80000){
      df$decision[i] <- "Admit"
    } else if(income < 40000 & c_record == FALSE){
      df$decision[i] <- "Admit"
    } else if(income < 40000 & c_record == TRUE){
      df$decision[i] <- "Admit"
    } else if(income >= 40000 & income <= 80000 & years > 3){
      df$decision[i] <- "Admit"
    }
  } 
  df
}

applicants_info <- data.frame(income = c(40000, 80000, 25000, 70000),
                              c_record = c(F, T, T, F),
                              years = c(2, 10, 3, 6),
                              stringsAsFactors = F) 


all_loan_offers(applicants_info)
#>   income c_record years decision
#> 1  40000    FALSE     2   Reject
#> 2  80000     TRUE    10    Admit
#> 3  25000     TRUE     3    Admit
#> 4  70000    FALSE     6    Admit

EDIT: Note that you can omit the for loop entirely, too. R is great because it operates on whole vectors of values. You can create numeric index from a logical vector and insert conformal vectors of values (thanks to "recycling" of the right hand side)

all_loan_offers <- function(df) {
  df$decision <- "Reject"
  df$decision[which(df$income > 80000)] <- "Admit"
  df$decision[which(df$income < 40000 & !df$c_record)] <- "Admit"
  df$decision[which(df$income >= 40000 & df$income <= 80000 & df$years > 3)] <- "Admit"
  df
}