Add random noise to POSIXct variable

Question

I am modelling a Poisson point process and the times in my data are in the POSIXct form and are only accurate to the second. Thus, there are some times that are the same. I want to add some noise to these times so hopefully, they can be different. Is there any package or function in R that allows me to do that?

If your vector is named `x` try to see if `format(x, "%Y-%m-%d %H:%M:%OS3")` are all equal. If they are, you can add `?runif`. Note: format `O3` prints with 3 decimals after the seconds, those decimals are already there, they simply aren't printed. — Rui Barradas, Jul 11 '20 at 12:10

score 1 · Accepted Answer · answered Jul 11 '20 at 12:22

As Rui points out, this is largely a matter of formatting. I think the simplest way to do this is to allow fractions of seconds to be printed when you are working with POSIXct - you can do this with:

 options(digits.secs = 3)

So now if I have a vector of times:

times <- as.POSIXct(c("2020-07-11 13:06:01", "2020-07-11 13:06:01"))
times
#> [1] "2020-07-11 13:06:01 GMT" "2020-07-11 13:06:01 GMT"

I can add fractions of seconds quite easily using the lubridate package:

library(lubridate)

times + seconds(runif(2))
#> [1] "2020-07-11 13:06:01.494 GMT" "2020-07-11 13:06:01.470 GMT"

In your case you probably want to add seconds(runif(length(times), -0.5, 0.5)) to keep your times randomized to within the nearest second.

Add random noise to POSIXct variable

1 Answers1