max identifier length
by Allin Cottrell
In gretl CVS I have now made a lot of small changes which
should permit an easy switch of the max legnth of names of
variables, governed by the definition of VNAMELEN in
lib/src/libgretl.h.
The changes may of course have introduced some errors, so I'd
be grateful if people could test things, in particular looking
for any broken printout.
The adventurous might like to try redefining VNAMELEN from 16
to 32 in libgretl.h and experimenting with longer names. In
this case you're almost certain to get some broken-looking
printout -- though model output, the output of "summary", and
that of "corr" should look OK, I think.
Allin
12 years, 2 months
note on sprintf command
by Allin Cottrell
Hello all,
This is something I've been vaguely aware of for a while, but now
(as I'm working on cleaning up gretl's internal handling of
user-defined variables) I've become acutely aware of it.
The "sprintf" command in gretl has a deplorable sloppiness. If you
do this
<hansl>
scalar x = 73
sprintf strvar "%g", x
</hansl>
gretl will not complain, and will create a string variable by the
name of "strvar" holding the string "73".
Further, if you do this
<hansl>
scalar x = 73
matrix strvar = I(3)
sprintf strvar "%g", x
</hansl>
gretl will still not complain, and again will add a string variable
called "strvar", which will never be seen since it is masked by a
pre-existing matrix variable of the same name.
Time for this madness to stop!
It won't be fixed in the forthcoming 1.9.10 release, but it will be
in 1.9.11: by that time, "sprintf" will expect its first argument to
name a pre-existing string variable into which to write its output,
and it will fail if there's no such variable.
<hansl>
# this is the right way!
scalar x = 73
string strvar
sprintf strvar "%g", x
print strvar
</hansl>
Allin Cottrell
12 years, 2 months
Strings for Translation
by Henrique Andrade
Dear Allin,
I'd found some untranslatable strings:
(1) "no valid values" when someone try to open
empty series
(2) "Bare declarations are not allowed here" when
someone try to use a foreach loop with strings
(e.g. loop foreach mes "1m" "2m" --quiet)
(3) "corr: o argumento 1 deve ser series, is matrix"
when someone try to calculate (in a wrong way)
the correlation of a matrix.
(4) "The variable X is read-only" when someone try
to rename a variable inside some kind of loops.
Could you please mark them for translation?
Best regards,
Henrique Andrade*
*
12 years, 2 months
generate dataset
by Sanzad Siddique
Hi,
I would like to generate an arma dataset using libgretl which is same as
the datase created by below code:
nulldata 2000
setobs 7 2006/06/02
series a = normal()
series z = 0
z = 0.4*z(-1) + 0.40*z(-2) + a + 0.5*a(-1) + 0.6*a(-2)
z1 = z-mean(z)
z2 = z/sd(z)
z3 = (z-mean(z))/sd(z)
I have started with DATASET *dset = create_new_dataset(3, 2000, 0); and
don't know how to move further. Is there any sample code for data series
generation like above? can anyone help?
Thanks in advance.
Regards,
Sanzad
12 years, 2 months
Improving Microdata Handling
by Henrique Andrade
Dear Gretl Team,
Lately, I started to study about microdata here in
Brazil. I confess this is a whole new field in eco-
nometrics for me.
Talking to another Brazilian econometricians I rea-
lized that we have basically three databases about
our population (all of them provided by the IBGE -
The Brazilian Institute of Geography and Statistics):
Censo (Population Census)
PNAD (National Household Sample Survey)
POF (Consumer Expenditure Survey)
I also realized that the great majority of the re-
searchers (not only econometricians, but also sta-
tisticians, pharmacologists, sociologists, etc.) uses
SPSS, Stata and SAS for data handling and poste-
rior analisys.
As a Gretl community (passionate!) member :-),
I started ask to myself: "Why not with Gretl?"
Due to this fact, I started to think about ways to
improve Gretl in order to gain a wider audience
among these potential users
The first step I am taking is to convert these da-
tasets into Gretl's native dataset. Hopefully, in
a couple of days I will start to report the first
results :-)
Finally, I would like to suggest some features that,
in my humble opinion, could improve Gretl (the
first two more directly related to microdata):
(1) Put string tables in the Icon View window;
(2) Allow a better fomating for the codebook (e.g.
enabling HTML or something like this);
(3) Allow description string for the scalars;
(4) Put variable lists in the Icon View window.
Many thanks in advance,
Henrique Andrade*
*
12 years, 2 months
Small bugs with invalid data series
by Sven Schreiber
Hi,
using a recent cvs build on Ubuntu:
Create a series with all NA's.
First glitch: Try to plot a boxplot, the error message says (slightly
misleading IMHO): "series is a constant"
Second glitch: Interpret data as time and series, and trying to do a
time series plot produces the following errors on the console:
<console output>
realgen: exiting on lex() error 15
gnuplot: using pngcairo driver
stderr: '
set xrange [1.797693135e+308:1.797693135e+308]
^
"/home/svenne/.gretl/gpttmp.OjFw46", line 18: undefined value
'
gretl_errmsg: '
set xrange [1.797693135e+308:1.797693135e+308]
^
"/home/svenne/.gretl/gpttmp.OjFw46", line 18: undefined value
'
Failed command: 'gnuplot "/home/svenne/.gretl/gpttmp.OjFw46"'
</console output>
thanks,
sven
12 years, 2 months
bug?
by Ignacio Diaz-Emparanza
When I run the attached script, it seems the function is correctly
generating the filtered series, and they are available from the command
line, but they don't appear in the main window, so I cannot make the
graphs from the menus. Some time ago this was different.
This is on Ubuntu linux 12.04 and with gretl 1.9.9cvs downloaded today,
but I also tried with a version from a month ago with the same problem.
So I think the question about the length of the names is not the problem
here.
--
Ignacio Diaz-Emparanza
DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
UPV/EHU Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
T.: +34 946013732 | F.: +34 946013754
www.ehu.es/ignacio.diaz-emparanza/
www.ea3.ehu.es
12 years, 2 months
Re: [Gretl-devel] libgretl arma
by Sanzad Siddique
Hi,
Can you please tell me, what value I should set for gretlopt if i want to
call arma function with conditional option. The default is ML exact (if i
pass OPT_NONE)
Also, can you add me (pavelkushtia(a)gmail.com) in the gretl-devel mailing
list.
Thanks,
Sanzad
12 years, 2 months
Re: [Gretl-devel] [ESS] Gretl mode for ESS
by Vitalie Spinu
>> ahmadou dicko <dicko.ahmadou(a)gmail.com>
>> on Mon, 1 Oct 2012 00:03:27 +0000 wrote:
ad> Thank's a lot Vitalie for your quick reply.
ad> I have never heard of gretl. How big is the community? And how does it
>> compare to R, ox and eviews?
>>
ad> Gretl is a software for econometric analysis ( also made by
ad> econometrician ) but you can do some statistical analysis too
ad> (even if I prefer R for that).
ad> Nowaday, It's widely known among econometrician and it's viable
ad> alternative to Stata and Eviews.
Yes, I have looked into it a bit closer. It's a pretty mature
kit. Detailed manual of 330 pages and very easy to use GUI. Not to
mention a huge set of scripts and datasets for Ramanathan, R. (2002) and
Greene, W. H. (20xx). It is indeed very focused and pretty much
complementary to R in terms of time series.
Help system is good. It will be very easy to integrate with ESS
help. Argument completion and eldoc are also pretty straightforward as
their help pages are well structured and easy to scrap (though it would
be good to have an internal function to retrieve arguments and command
options directly).
Gretl has a pretty nice menu interface. I wonder if there is a way to
start a GUI dialog directly from gretl console. Would be nice to
automatically generate all gretl menus in emacs such that users familiar
with GUI can feel at home.
I am pretty convinced now that gretl would be a nice addition to ESS,
especially in the light of it being GNU licensed.
Vitalie
ad> I think that R and Gretl can play nice together because Gretl is
ad> really fast (really fast) and have some nice time series library
ad> (I just started to write a R package to exchange easily data
ad> between R and Gretl https://github.com/dickoa/gretlReadWrite)
ad> Comparing to R and I think that the community is smaller but It
ad> growing very fast (active mailing list, yearly
ad> conference<http://www.gretlconference.org/>and some book about the
ad> software).
>> It's not that difficult to add a new interaction mode to ESS. The main
>> problem is not with adding a mode, but rather with maintaining it
>> latter on.
>>
>> I can look into it, but not earlier than in a couple of months time.
>>
ad> Great !!!
>> You can look into it yourself. Sources for inspiration are
>> ess-julia.el, which you probably should start with, and also
>> ess-r-d.el for a more complete picture. Once the editing mode is
>> there, the main step is to construct ess-gretl-customize-alist with
>> all the defaults needed for the interaction with the subprocess,
>> and then write 'gretl function for starting a subprocess.
>>
ad> Thank's a lot...ESS 12.09 is very nice and I already tried to play
ad> with Julia (which is really interesting too). So, I'll look into
ad> those files and try to understand how their works
ad> Thank's again
ad> On Sun, Sep 30, 2012 at 10:00 PM, Vitalie Spinu <spinuvit(a)gmail.com> wrote:
>> >> ahmadou dicko <dicko.ahmadou(a)gmail.com>
>> >> on Sun, 30 Sep 2012 13:04:40 +0000 wrote:
>>
>> > Hi everyone,
>> > I would like to know if adding a GRETL <http://gretl.sourceforge.net/>> mode
>> > to ESS is planned ?
>>
>> I have never heard of gretl. How big is the community? And how does it
>> compare to R, ox and eviews?
>>
>> > There's already a basic gretl
>> > mode<
>> http://fossies.org/unix/misc/gretl-1.9.9.tar.gz:a/gretl-1.9.9/utils/emacs...> >inemacs
>> > but it lacks several features availables in ESS R-mode (sending
>> > command line, buffer, autocompletion, etc)
>>
>> It's not that difficult to add a new interaction mode to ESS. The main
>> problem is not with adding a mode, but rather with maintaining it latter
>> on.
>>
>> I can look into it, but not earlier than in a couple of months time.
>>
>> > So my question is if it's not planned, where to start ? (I would like
>> to
>> > help even if my elisp programming skillz is weak)
>>
>> You can look into it yourself. Sources for inspiration are ess-julia.el,
>> which you probably should start with, and also ess-r-d.el for a more
>> complete picture. Once the editing mode is there, the main step is to
>> construct ess-gretl-customize-alist with all the defaults needed for the
>> interaction with the subprocess, and then write 'gretl function for
>> starting
>> a subprocess.
>>
>>
>> Vitalie
>>
12 years, 2 months
Gretl project on the launchpad
by Ivan Sopov
Hello, gretl developers.
I'm trying to start a translation of help files into russian on
launchpad.net as it seems to be the most suitable tool to participate
for all familiars with econometrics but not with gettext, linux. cvs,
etc.
The problem is that there is already a project for gretl on launchpad
and it is strongly prohibited to start more than one project for a
single program. I cannot contact with Constantine Tsardounis for about
a month, so I think it is time to re-assign that project to someone
else. On the irc-channel of launchpad I was told that
Our admins can re-assign the project to new owners but we'd prefer to
hear from the upstream owners. can you get one of them to submit a
question here:
https://answers.edge.launchpad.net/launchpad
But if nobody from main developers wants to register and do something
at launchpad it is possible to assign this function to me and in that
case a letter in this list will probably be enough.
I have prepared a .po-file for genr_funcs.xml and gretl_commands.xml
with the help of po4a utility and got 1511 strings for translation
(strings a rather big).
Good luck, Ivan Sopov.
P.S. My previous letter about using launchpad for translation is
http://lists.wfu.edu/pipermail/gretl-devel/2009-November/002171.html
12 years, 3 months
