Specifying column with its index rather than name

Question

I have written a function to get the Proportional Stacked Bar plot using ggplot function. Right now I am using Column name in this ID .

PropBarPlot<-function(df, mytitle=""){
    melteddf<-melt(df, id="ID", na.rm=T)
    ggplot(melteddf, aes(ID, value, fill=variable)) + 
      geom_bar(position="fill") + 
      theme(axis.text.x = element_text(angle=90, vjust=1)) + 
      labs(title=mytitle)
}

I want to make it generic. So I want to make use of column index instead of column name. I tried doing something like this.

PropBarPlot<-function(df, mytitle=""){
    melteddf<-melt(df, id=names(df)[1], na.rm=T)
    ggplot(melteddf, aes(names(df)[1], value, fill=variable)) + 
      geom_bar(position="fill") + 
      theme(axis.text.x = element_text(angle=90, vjust=1)) + 
      labs(title=mytitle)
}

But of no use. Can someone suggest me how to do this??

Thanks.

Answer 1

As pointed out by @baptiste you should use aes_string() instead of aes() to use strings in defining x and y values. Also you should put value and variable inside quotes.

PropBarPlot<-function(df, mytitle=""){
  melteddf<-melt(df, id=names(df)[1], na.rm=T)
  ggplot(melteddf, aes_string(x=names(df)[1],y= "value", fill="variable")) + 
    geom_bar(position="fill") + 
    theme(axis.text.x = element_text(angle=90, vjust=1)) + 
    labs(title=mytitle)
}

Answer 2

As pointed out by @Bryan Shalloway, the aes_string() approach is now soft deprecated, and replaced by tidy evaluation. Using the tidy evaluation approach, my solution would be:

library(reshape)
library(ggplot2)

# let's start by creating an example dataframe
id <- c(1, 2, 3, 4)
var1 <- c(10, 20, 10, 20)
var2 <- c(6, 3, 2, 5)
df <- data.frame(id, var1, var2)

# Now let's build the function
PropBarPlot<-function(df, mytitle=""){
  # here I create a variable that contains the first element of the vector 
  # of df column names (it's a string)
  my_name <- colnames(df)[1]
  # here we melt, using the new variable as the id parameter
  melteddf<-melt(df, id=my_name, na.rm=T)
  # and here we plot, using the .data pronoun and the new variable
  ggplot(melteddf, aes(x = .data[[my_name]],y = value, fill = variable)) + 
    geom_bar(position="fill", stat="identity") + 
    theme(axis.text.x = element_text(angle=90, vjust=1)) + 
    labs(title=mytitle)
}

If you would like to learn more about tidy evaluation (with step-by-step explanation and practical examples), I wholeheartedly recommend the learnr tutorial by Ian Lyttle, available at https://ijlyttle.shinyapps.io/tidyeval/