Generating a moving sum variable in R

Question

I suspect this is a somewhat simple question with multiple solutions, but I\'m still a bit of a novice in R and an exhaustive search didn\'t yield answers that spoke well to wha

Answer 1

you can also use filter of standard packages (stats) to do moving sum:

ms=function(x,n=5) as.numeric(stats::filter(x,rep(1, n),method="convolution",sides=1))
x=c(1,2,3,4,5,6,7,8,9)
ms(x,5)
NA NA NA NA 15 20 25 30 35

To do a 1-lag, insert NA at the begining and take the number of elements (or lines):

ms.1lag=c(NA,ms(x,5))[1:length(x)]
cbind(x,ms.1lag)
x ms.1lag
[1,] 1      NA
[2,] 2      NA
[3,] 3      NA
[4,] 4      NA
[5,] 5      NA
[6,] 6      15
[7,] 7      20
[8,] 8      25
[9,] 9      30

If you use this frequently,

ms=function(x,n=5,lag=0)
  c(rep(NA,lag),
    as.numeric(stats::filter(x,rep(1, n),method="convolution",sides=1)))[1:length(x)]
cbind(x,ms5.1=ms(x,5,1))
      x ms5.1
 [1,] 1    NA
 [2,] 2    NA
 [3,] 3    NA
 [4,] 4    NA
 [5,] 5    NA
 [6,] 6    15
 [7,] 7    20
 [8,] 8    25
 [9,] 9    30