R在处理时间序列（ts）对象而不是data.frame时出错

Question

I want to use a time series object (because I want to use lag() later) but I am not able to adress the time series object in a function:

With a data.frame it does not matter whether I code

dat=data.frame(x=c(1,2,3,4,5,6))
sum(2*dat[,"x"])

[1] 42

or alternatively

with(dat,sum(2*x))

[1] 42

But as soon as I transform the data.frame into a time series object the function does not work anymore.

dat=data.frame(x=c(1,2,3,4,5,6))
dat <- ts(dat)
sum(2*dat[,"x"])

[1] 42

so this still works, but

with(dat,sum(2*x))

now results in an

Error in eval(substitute(expr), data, enclos = parent.frame()) : 
  numeric 'envir' arg not of length one

Or to put it in other words: With the ts-object I cannot use any functions like

testy <- function(data,par){
  with(data,sum(par * x))
}
sapply(data=dat,2,testy)

without getting that error "numeric 'envir' arg not of length one" whereas df-objects do not cause that error. But I need ts() later for the lag() function.

What can I do to use time series objects within a function?

Answer 1

We could convert it to zoo and it has with as one of its methods

library(zoo)
methods(class = 'zoo')
with(zoo(dat), sum(2 * x))
# [1] 42