r - using case_when directly on a factor

Question

I can't figure something straightforward to do this simple thing :

I have a variable age_cl which is a factor object with age as classes.

If I want to merge two levels I'll just go easily with a forcats::fct_collapse.

However, now I would like to split one level in two different levels (e.g. the "25-34" level into a "25-29" and a "30-34" using the initial integer age variable to conditionally split).

I couldn't find any equivalent in forcats to do so, hence I'd like to use a standard case_when but it only takes character vector as input...

Anyone with any thoughts on this ?

Thanks a lot

question from:https://stackoverflow.com/questions/65952962/using-case-when-directly-on-a-factor

Answer 1

Have you considered cut ? You can put custom breaks and labels for it.

breaks <- c(24, 29, 34)
labels <- paste(breaks[-length(breaks)] + 1, breaks[-1], sep = '-')
labels
#[1] "25-29" "30-34"

transform(df, age_cl = cut(age, breaks, labels))

#   age age_cl age_cl_expected
#1   25  25-29           25-29
#2   26  25-29           25-29
#3   27  25-29           25-29
#4   28  25-29           25-29
#5   29  25-29           25-29
#6   30  30-34           30-34
#7   31  30-34           30-34
#8   32  30-34           30-34
#9   33  30-34           30-34
#10  34  30-34           30-34