Re: [Gretl-devel] 'series x' generate zeroes?
Am 07.06.2011 09:48, schrieb Riccardo (Jack) Lucchetti:
On Fri, 3 Jun 2011, Ignacio Diaz-Emparanza wrote:
> ¿Is it normal that the simple command
>
> series x
>
> generates zeros ? I thought it should generate NAs
Hm, you may be right. However: is there any particular reason why one
choice should be strongly preferred to the other?
Yes, because zeroes look like valid data, which in this case they are
not. Imagine that further down you do something with x in a restricted
sample, and then revert to the full sample again.
There is a strange kind of zero affinity among the gretl coders it seems
:-) (think of 0*NA==0)
-sven