[Gretl-devel] Re: time series cmds like corrgm in panel data
On Tue, 4 Nov 2025, Sven Schreiber wrote:
And now I'm also seeing a complementing problem, where lrvar
doesn't
work in a panel environment although it should (IMHO):
<hansl>
open grunfeld
eval lrvar(mnormal(200)) # error
</hansl>
I think lrvar should be allowed to work on a vector/matrix no matter
whether there's currently a panel dataset in place.
Yes, agreed. Given a vector argument we should assume the caller knows
what he's doing (in particular, that he's not passing a panel series
in the guise of a vector, since lrvar doesn't handle panel data), and
that's now the case in git.
And what's more, I saw this behavior inside a function which
didn't
get any series context as arguments - only the caller lived in a
panel.
By design, a gretl function always gets a pointer to the caller's
dataset. Otherwise it would be impossible to create a series inside a
function unless it calls for a series or list argument. So that
observation is not really valid "more". We should ignore the character
of the dataset when given vector input, is all.
Allin