Re: [Gretl-devel] The mysterious stack() function
Am 10.09.2018 um 02:30 schrieb Allin Cottrell:
And I thereby left the business unresolved! Here's an attempt at
resolution. It turns out that the problem for which stack() provides a
solution is pretty easily handled by
series -> matrix -> reshape matrix -> series
...
So here's a suggestion: in section 4.5 of the Guide I delete the
discussion of the fix using the mysterious stack() in favour of an
example and explanation of the method illustrated above.
Should I do that? If so, stack would then be undocumented, but we
could keep it for the time being for backward compatibility.
Thanks for the suggestion. What I like about it is to clean up the gretl
syntax and remove that "oddball", as you called it. However, replacing a
built-in command/function/whatever by a description of a relatively more
advanced scripting solution is not optimal IMO.
First I thought the functionality of stack() (and of your solution)
could be merged into the 'dataset' or the 'setobs' command -- but
actually the description in the guide ch.4 and the whole business of
stack() is to read in data from external files. So wouldn't 'join' be a
natural candidate for this stuff?
So from a user's point of view the first stack() example in section 4.5
would become something like this:
join panel.txt x1 --panelstack --length=50
join panel.txt x2 --panelstack --length=50 --offset=50
(Perhaps a --no-header option might be needed, but OTOH it might be
implicit in --panelstack.)
I agree that stack() should be kept for a while. For me personally it
was undocumented BTW, because I rely heavily on the built-in command and
function references, where it doesn't appear. Hence the "mysterious" in
this thread's title.
thanks,
sven