On Thu, 4 Oct 2007, Ignacio Diaz-Emparanza wrote:
> > I think the 'smpl' command maybe used in the form
you need. For example,
> > to work with all observations except 3,5 and 11-22 you may use
> >
> > smpl (obs<11 || obs>22) && obs!=3 && obs!=5 --restrict
>
> But I want to delete the observations, not just restrict the
> analysis to some observations.
Then, as Ignacio says, all you have to do is save the sub-sampled
dataset.
If the dataset is sub-sampled on a boolean criterion, the "store"
command automatically saves the subsample.
In the GUI, if you sub-sample then go to save the data, you're
prompted as to whether you want to restore the full sample first.
If you say No, you're then asked if you want to switch to the
reduced dataset (in effect, clear the old dataset and load the
newly saved version).
Allin.