Create New Columns Within a Loop

It sounds like you are trying to create dummy variables using your data.

Note that for most modeling functions in R this will happen automatically in the modeling step. The way this works is using the model.matrix() function.

Here is an example:

df <- data.frame(age=sample(1:100,10,replace=T))

# Create a categorical variable using cut()
df$agegroup <- cut(df$age, breaks=seq(0, 100, by = 10))

You now have a categorical variable with age groups:

head(df)
  age agegroup
1  82  (80,90]
2  79  (70,80]
3  99 (90,100]
4  12  (10,20]
5  82  (80,90]
6  66  (60,70]

Convert to a model matrix

# Create the model matrix

model.matrix(~agegroup - 1, df)
   agegroup(0,10] agegroup(10,20] agegroup(20,30] agegroup(30,40] agegroup(40,50]
1               0               0               0               0               0
2               0               0               0               0               0
3               0               0               0               0               0
4               0               1               0               0               0
5               0               0               0               0               0
6               0               0               0               0               0
7               0               0               0               0               0
8               0               1               0               0               0
9               0               0               0               0               1
10              0               0               0               0               0

Use <- instead of <<-. Using <<- is assigning things in the global scope, where d does not exist. Finally, return d.

add_ages <- function(d) {
  for (i in seq(10,100,10)){
    d[,paste0("age",i)] <- ifelse(i>=d[,"age"] & d[,"age"]<i+10,1,0)
  }
  d
}
df <- add_ages(df)

Edit:

If you really want to avoid doing df <- add_ages(df), you could do the following:

add_ages <- function() {
  for (i in seq(10,100,10)){
    df[,paste0("age",i)] <<- ifelse(i>=df[,"age"] & df[,"age"]<i+10,1,0)
  }
}

add_ages()

I'd recommend against this for at least two reasons. First, this does not generalize at all. There's really no point in making a function that does this, you'd be better off just using the loop immediately after creation of df, e.g.

df <- data.frame(age=sample(1:100,10,replace=T))
for (i in seq(10,100,10)){
  df[,paste0("age",i)] <<- ifelse(i>=df[,"age"] & df[,"age"]<i+10,1,0)
}

Second, functions should strive to avoid side effects. In other words, if I call a function, the only object that gets modified is where I save the output to. Side effects like this may seem harmless, but if you were to write this as one of several functions in the middle of some code, and then come back to it 6 months later, its likely you'll forget side effects, which can cause all sorts of headaches.