Assign value to dynamic column name in a function in R-CodePudding

I have a table:

num1<-runif(10)
num2<-runif(10)
df<-data.frame(num1,num2)
df
         num1      num2
1  0.41170486 0.8198323
2  0.46594131 0.7300186
3  0.09005351 0.9667960
4  0.33373968 0.6827277
5  0.48305910 0.2254627
6  0.23514394 0.3348283
7  0.79450479 0.3101271
8  0.30476853 0.6706657
9  0.53643541 0.9761387
10 0.66568823 0.8444414

And I want to create a percentile column for each of these 2 columns respectively so the table looks like this:

 df
         num1      num2.   num1_percentile   num2_percentile
1  0.41170486 0.8198323    20%                40%
2  0.46594131 0.7300186    xxx               xxx
3  0.09005351 0.9667960    xxx               xxx
4  0.33373968 0.6827277    xxx               xxx
5  0.48305910 0.2254627    xxx               xxx
6  0.23514394 0.3348283    xxx               xxx
7  0.79450479 0.3101271    xxx               xxx
8  0.30476853 0.6706657    xxx               xxx
9  0.53643541 0.9761387    xxx               xxx
10 0.66568823 0.8444414    xxx               xxx

So the percentile column name will be based on the if it's num1 or num2. I have this function where I am trying to assign percentile values to dynamic column names but seems like assign() did not store values in num1_percentile.

create_percentile<-function(num){
  
  breaks=c( quantile(df[[num]], probs =seq(0, 1, 0.2)))
  
  
  
  values<- cut(df[[num]],
               breaks=breaks,labels=c(  '20%','40%', '60%','80%','100%'))
  values[is.na(values)]<-'20%'
  
  print(values)
  
  assign(paste0(df,'_percentile'),values)
  
  df<-cbind(df,paste0(num,'_percentile'))
  
  return(df)
}

create_percentile('num1')

How do I make it work so it can assign values to dynamic column names?

CodePudding user response：

The reason behind assign not storing the value is that you are storing it in assign(paste0(df,'_percentile'),values) where df is a data frame. So while assigning it considers the first element of df to paste it with _percentile. A small correction is instead of df you can use num.

Also this function cbind(df,paste0(num,'_percentile')) won't work as expected because it will store a character value returned from paste command.

You may use the below code to assign the values to dynamic column names

num1 <- runif(10)
num2 <- runif(10)
df <- data.frame(num1, num2)
df

         num1       num2
1  0.78584180 0.61999953
2  0.33916132 0.91999116
3  0.84080205 0.19279828
4  0.03563479 0.04253741
5  0.16633144 0.23509481
6  0.15459396 0.84458053
7  0.99544676 0.08727064
8  0.25060605 0.51908750
9  0.35631173 0.48853416
10 0.50649402 0.64245468

create_percentile <- function(num) {
  breaks <- quantile(df[[num]], probs = seq(0, 1, 0.2))
  values <- cut(df[[num]], breaks = breaks, labels = c('20%', '40%', '60%', '80%', '100%'))
  values[is.na(values)] <- '20%'
  df <- cbind(df, values)
  names(df)[ncol(df)] <- paste0(num, '_percentile')
  return(df)
}

df <- create_percentile('num1')
df <- create_percentile('num2')

Output:

         num1       num2 num1_percentile num2_percentile
1  0.78584180 0.61999953             80%             80%
2  0.33916132 0.91999116             60%            100%
3  0.84080205 0.19279828            100%             40%
4  0.03563479 0.04253741             20%             20%
5  0.16633144 0.23509481             40%             40%
6  0.15459396 0.84458053             20%            100%
7  0.99544676 0.08727064            100%             20%
8  0.25060605 0.51908750             40%             60%
9  0.35631173 0.48853416             60%             60%
10 0.50649402 0.64245468             80%             80%