Simulating from a vector of discrete data

Question

I have a vector of discrete data and I want to simulate from the empirical distribution associated to this data, I was simulating with the function rlogspline after doing fit<-logspline(vector_of_data) where vector_of_data is data that is suppose to be coming from a continuous distribution, that's why I used logspline, but with this vector I have the certainty that the values in it are of discrete nature so I can't use logspline to adjust a "fit" for it.

Basically what I want to do is to adjust a "fit" of the observed data and then use that fit to simulate those values. Do you think this can be done in R?

Thank you very much for your help.

Answer 1

I am not totally clear exactly what you are trying to do, but could you use something like quantile and runif, for example:

obs <- c(125,110,115,100,150)             # original observations
sim <- quantile(obs, runif(10000))        # simulations
hist(sim, freq=FALSE)

Answer 2

I think sample(x,...,replace=TRUE) (sampling with replacement) should simulate from the empirical distribution ...