[Gretl-users] Re: PWT data (was Re: Re: [EXTERNAL] Re: panplot question)
I'll pick one of Sven's points for now...
On Sat, Jan 28, 2023 at 12:54 PM Sven Schreiber
<sven.schreiber(a)fu-berlin.de> wrote:
Why actually are the ECB's AWM datasets located in the sample datasets collection
instead of the databases? They're not textbook-style illustration data. Or am I
misunderstanding the rules that determine where a dataset/database is to be found in
gretl? Even after all these years I find it quite confusing, actually.
* If a data collection contains series of different frequencies then
(leaving aside the special case of MIDAS files) it must be packaged as
a gretl database.
* It may also be preferable (more parsimonious) to use the gretl
database format if a collection contains series with substantially
different observation ranges. In a regular gretl datafile all series
must be of a common length (padded with NAs if need be).
* There's a certain presumption in favour of a database for very large
collections of series.
That leaves some room for discretion. I guess we could go either way
with the AWM files but they're quite compact so using regular datafile
format seems OK. The supplied datafiles under the "gretl" heading are
mostly not textbook-oriented.
Allin
--
Allin Cottrell
Department of Economics
Wake Forest University