R bootstrap time series

Question

I would like to generate four bootstrap samples from a time series data set and have each new bootstrapped sample become a new list element. The sample size needs to be the same length as the original data set. Can someone please give me a hand? This is all I could come up with so far.

data <- ts(matrix(rnorm(36), 12, 3), start=c(2012, 1), frequency=12)
data
replicate(4, apply(data, 1, sample, replace=TRUE))

Answer 1

You were nearly there. Your description wasn't totally clear at first, but the comment cleared it up. You need to apply across columns, and use list to make each replicate a list element:

boot <- replicate( 4 , list( apply(data , 2 , function(x) sample( x , replace=TRUE ) ) ) )

class(boot)
#[1] "list"

length( boot )
#[1] 4

head(boot[[1]])

#       Series 1    Series 2   Series 3
#[1,]  0.4652513 -0.02065698  0.3328945
#[2,]  0.6649865  0.08845410  0.2032134
#[3,]  0.5975473 -1.64571306  1.6516726
#[4,]  0.5975473 -0.23359075 -0.3255437
#[5,]  0.4008458  0.42180633  1.8402009
#[6,] -0.5436319  1.17034910  0.3456304

EDIT

Since you need to do whole rows at a time this is even easier!

boot <- replicate( 4 , list( data[ sample( nrow(data) , replace = TRUE ) , ] ) )