Question

This problem has plagued me for quite some time. I always just work around it with a for loop, but I think it’s finally time for me to find a quicker and more elegant way of doing this.

As an example, let’s say I have a data frame containing information on whether an item is red or blue. The information is presented in this way:

item.df <- data.frame(Item=seq(1,5), Red=c("Y", "Y", "N", "N", "N"), Blue=c("N", "N", "Y", "Y", "N"))

Clearly, this is not the most condensed way to represent this information. Instead of having two separate Red and Blue columns, I simply want one item color column that would contain “Red”, “Blue”, or “Neither” (Or NA would also be acceptable).

Obviously, I can achieve this by creating an empty Item.Color column and then filling it in by looping through each individual row. But I’m sure there is a quicker way to do this.

Back when I was a true R novice, I tried to do it by:

item.df$Item.Color <- if(item.df$Red=="Y"){"Red"}

but I quickly learned this doesn’t work, because the if statement will only read the first element in item.df$Red.

Could there be a way to achieve this using do.call() or one of the apply() functions? I’ve attempted, but I could never get it to do quite what I wanted. Thanks in advance for any insight you might be able to provide!

p.s. I would also be grateful to hear any suggestions for a better title for this question. For me, that always seems to be the hardest part in asking questions.

Answer 1

The following code should do the trick, it even checks if the data contains rows where both Red and Blue are TRUE (== "Y").

item.df[["condensed"]] <- NA
item.df = within(item.df, {
  if(any(Red == "Y" & Blue == "Y")) stop("Blue and red cannot both be TRUE")
  condensed[Red == "Y"] <- "Red"
  condensed[Blue == "Y"] <- "Blue"
})

The trick here is that the same syntax for taking a subset can be used for assignment:

# Getting a subset from a vector
A[B > 2]
# Or assigning NA to that subset 
A[B > 2] <- NA