Rollapply() backwards in R

问题:

Is there a way to apply rollapply() backwards say from 100 to 1 instead of 1 to 100 or should I sort my data first and then apply rollaplly()?

回答1:

Try this

z = 1:100
#Normal rollapply
rollapply(z, 2, mean)

#Reverese
rollapply(z[length(z):1], 2, mean)