On Sat, 10 Sep 2011, artur bala wrote:
 Le 10/09/2011 12:15, Sven Schreiber a écrit :
> Yes, but he could do the resampling twice and take the first 200 obs
> from the first series, and the final 50 obs from the second series,
> and glue the parts together to get a 250-obs series.
 Is this what you mean Sven ?
 nulldata 200
 series x = normal()
 # resampling twice
 series y1 = resample(x)
 series y2 = resample(x)
 # adding observations
 addobs 50
 # stacking y2 to y1
 loop i = 201..250 --quiet
     y1[i] = y2[i-200]
 endloop
 # here's the final series y
 series y = y1
 delete y1 y2
 print x y --byobs 
The alternative, if you don't mind resampling the whole dataset, is
dataset resample N
This gives you N draws from the original data, where N can be less 
than, equal to, or greater than the original number of observations.
Allin Cottrell