On Thu, 24 Nov 2011, Federico Lampis wrote: > have a question about the use the "smpl full" inside a function > suppose this case: > > nulldata 50 > smpl 1 40 > > smpl full > series t1=time > scalar nt=lastobs(t1) > > in a script these commands produce: > > Generated scalar nt = 50 > > But if I write something like this: > > nulldata 50 > smpl 1 40 > > function scalar example (----) > > smpl full > series t1=time > scalar nt=lastobs(t1) > end function > > the scalar is nt=40. The question: there is some trick to recover the > full sample inside a function or the last value of the original full > sample without specify any argument about it in the function > declaration? The rule is that a function only has access to the data range that was passed in by the caller. You can subsample within a function but smpl --full just takes you back to the dataset state as the caller specified it. If you want a function to do something with a sub-sample and also to do something with the full dataset, you can pass in the full dataset (that is, do "smpl --full" before calling the function) but pass in the sub-sample information as argument(s). For example int arguments t1 and t2, or a dummy series. Allin Cottrell _______________________________________________ Gretl-devel mailing list Gretl-devel(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-devel