optimized rolling functions on irregular time series with time-based window

Question

Is there some way to use rollapply (from zoo package or something similar) optimized functions (rollmean, rollmedian etc) to compute r

Answer 1

I recommend using runner package which is optimized to do operation requested in this topic. Go to section Windows depending on date in documentation, for further explanation.

To solve your task, one can use runner function which can execute any R function in running windows. One-liner here:

df <- read.table(
  text = "date  value
   2011-11-01      5
   2011-11-01      4
   2011-11-01      2
   2011-11-08      1
   2011-11-13      0
   2011-11-14      0
   2011-11-15      0
   2011-11-18      1
   2011-11-21      4
   2011-12-05      3", header = TRUE, colClasses = c("Date", "integer"))

library(runner)
runner(df$value, k = 5, idx = df$date, f = median)
[1] 5.0 4.5 4.0 1.0 0.0 0.0 0.0 0.0 2.5 3.0

---

P.S. one should be aware that 5-days window is [i-4, i-3, i-2, i-1, i] instead of (i-5):i (6-days window). Illustration below for better explanation of the concept.
I've made example on 5-days window but if one want to reproduce result as OP requested, can specify 6-days window:

identical(
  runner(df$value, k = 6, idx = df$date, f = median),
  c(5.0, 4.5, 4.0, 1.0, 0.5, 0.0, 0.0, 0.0, 2.5, 3.0)
)
# [1] TRUE