Question

I’m generating a large amount of lagged variables (365, actually) for a time series with 79,000 observations.

I currently have a column in a dataframe OrdersData called prospectdrops. I first convert it to zoo format to use the lag() function:

prospectdrops<-zoo(OrdersData$prospectdrop)

then I execute the for loop:

for (i in 1:365){
  prospectdrops[paste("lag",i,"day",sep="")] <- lag(prospectdrops,i*24,na.pad=TRUE)
}

then I’m forced to loop again using cbind to bind them to prepare them for data.frame()

for (i in 1:365){
cbind(prospectdrops, prospectdrops[paste("lag",i,"day",sep="")])
}

Understandably, this takes forever with the for loop function in R. I know that “apply” could be the answer, but don’t see a direct comparator in the descriptions of the functions. Any ideas?

Answer 1

k in lag.zoo can be a vector. See ?lag.zoo.

x <- zoo(11:21)
lag(x,1:3)