Re: [Gretl-users] Please test! (was Re: gretl 1.6.4)
by a b
Allin Cottrell a écrit :
On Fri, 18 May 2007, a b wrote:
It seems that the problem lies with the X-Y axis
combination...
....I believe this is fixed in CVS and the current snapshot.
It works fine now. Thanks ! :-)
Artur
---------------------------------
Stockage illimité de vos mails avec Yahoo! Mail. Changez aujourd'hui de mail !
16 years, 11 months
Re: [Gretl-users] Please test! (was Re: gretl 1.6.4)
by a b
Allin Cottrell a écrit :
On Thu, 17 May 2007, a b wrote:
"'" in variable's display name that still causes gnuplot to
crash -- supposed to have been fixed !
I can't replicate that problem any more. If you still see it with
1.6.5 please post a precise recipe for getting the gnuplot
failure. Thanks.
Ok! After having done some tests with 1.6.5, here's what I come out with:
It seems that the problem lies with the X-Y axis combination.
Suppose you have two variables, X & Y (undated or time series)
1. I insert a single quote ( ' ) into X display name and execute "gnuplot" command twice.
"gnuplot X Y" --is OK but "gnuplot Y X" --crashes.
2. I insert a single quote into Y display name too. Both variables have ( ' ) in their display names now.
In such a case both plots are OK: "gnuplot Y X" --is OK, "gnuplot X Y" --is OK.
3. I keep only Y with ( ' ) in its display name. Then "gnuplot Y X" --is OK but "gnuplot X Y" --crashes.
So, if only the X-axis variable contains the single quote, then gnuplot crashes.
By the way, would it be more intuitive that the gnuplot's arguments be "xvars yvars" instead of the actual ?
PS. The userdir works fine now!
Thanks,
Artur
---------------------------------
Stockage illimité de vos mails avec Yahoo! Mail. Changez aujourd'hui de mail !
16 years, 11 months
Re: [Gretl-users] Please test! (was Re: gretl 1.6.4)
by a b
Allin Cottrell a écrit :
Now: please beat on the Windows snapshot and see if it's OK!
Sorry to be late!
As for me, this last version is OK. Excepting the usedir feature that doesn't work and the single quote
"'" in variable's display name that still causes gnuplot to crash -- supposed to have been fixed !
Thanks,
Artur
---------------------------------
Stockage illimité de vos mails avec Yahoo! Mail. Changez aujourd'hui de mail !
16 years, 11 months
Re: [Gretl-users] seasonal plot
by a b
Marco Marini a écrit : Hi everybody,
for seasonal time series it is often useful to graph
its observations period-by-period (every January, every February,...).
It can be easily programmed in a script, but I wonder if it is possible to
insert a specific command in the menu.
Thanks!
Marco Ciao Marco!
I agree on the utility of a by-period graph but I guess it will be rather tricky to implement. But I hardly see how it can be "easily" programmed in a script since the "gnuplot" command does produce a file and not an image on the screen. Moreover, this command is not available in a loop. Alas!
Instead, your message instilled in me a reflection. I think, it would be a good starting point if one could display variables on a by-period basis after having restricted the sample. I mean: actually, if you have a quarterly data file and set your sample on january, when displaying the variables the date disappears, replaced by "Obs 1, Obs 2 ....". The date is recovered only by setting the full range.
Distinti saluti, :-)
Artur
---------------------------------
Découvrez une nouvelle façon d'obtenir des réponses à toutes vos questions ! Profitez des connaissances, des opinions et des expériences des internautes sur Yahoo! Questions/Réponses.
16 years, 11 months
gretl 1.6.5 plans
by Allin Cottrell
Just to let you know: I'll be in transit for the next 24 hours ro
so. I will try to get 1.6.5 properly released by Friday. If
anyone wants to submit updated translations before then I'll be
glad to include them. (There are a few new strings associated
with the new /Tools/Nonparametric tests items.)
Allin.
16 years, 11 months
gretl 1.6.4
by Allin Cottrell
Apologies to those who receive double notification -- and I won't
do this again, because I'll post right now the URL for the
gretl-announce list
http://ricardo.ecn.wfu.edu/gretl_announce.html
but, for anyone who's not on the announce list, gretl 1.6.4 is now
available at http://gretl.sourceforge.net/
This is mostly a quick fix for breakage in the Windows version of
gretl.
--
Allin Cottrell
Department of Economics
Wake Forest University, NC
16 years, 11 months
Ang. Re: Ang. [Gretl-users] Please test! (was Re: gretl 1.6.4)
by andreas.karlsson＠ltv.se
cottrell(a)wfu.edu @ INTERNET skrev 2007-05-15 18:20:46 :
> On Tue, 15 May 2007, andreas.karlsson(a)ltv.se wrote:
>
> > Thanks a lot. I see that you also have added the runs test in the Tools
>
> > Nonparametric tests > Runs test menu. That is great. But it still
appears
> > in the Variable menu. Maybe it should be removed from the Variable menu
so
> > it does not appear twice in different menus?
> > > Secondly, I see that the --verbose (print extra output) command can
not be
> > chosen from the Tools > Nonparametric tests > Difference test menu.
Could
> > you please add a check box for this command?
> > > Thirdly, I see that there are no Help buttons for the new Runs test
or
> > Difference test menus. Could these be created by slightly modifying the
> > runs and difftest command syntax help files?
>
> These things are now done.
Thank you very much.
Andreas
16 years, 11 months
seasonal plot
by Marco Marini
Hi everybody,
for seasonal time series it is often useful to graph
its observations period-by-period (every January, every February,...).
It can be easily programmed in a script, but I wonder if it is possible to
insert a specific command in the menu.
Thanks!
Marco
*-------
Marco Marini*
ISTAT - National Statistical Institute
Quarterly National Accounts (DBS/D)
Via Agostino Depretis 74/b
00184 Rome
+39.06.4673.3209
marco.marini(a)istat.it
www.istat.it
-----
16 years, 11 months
new py4gretl package for common-trends tests
by Sven Schreiber
Hi,
I have uploaded a new function package in the py4gretl family,
"py4gretl_ct_test". With it you can test a multivariate time series for
the number of stochastic trends as suggested by Nyblom/Harvey (kind of
like a multivariate KPSS test). It also incorporates the extensions by
Busetti for several types of deterministic breaks in the data.
As always, any feedback is appreciated.
-sven
P.S. This is really an example of something that could be done in native
gretl script quite easily I guess -- But since I had the numpy code
lying around...
16 years, 11 months
Ang. [Gretl-users] Please test! (was Re: gretl 1.6.4)
by andreas.karlsson＠ltv.se
cottrell(a)wfu.edu @ INTERNET skrev 2007-05-15 00:46:38 :
> On Mon, 14 May 2007, Sven Schreiber wrote:
>
> [ Should we release 1.6.5 real soon? ]
>
> > I don't know what the release policy is, but I would tend to answer
yes;
> > the user dir problem should mostly affect existing gretl users, who
will
> > be quite puzzled why they can't get their old settings to work.
>
> Yes, agreed. I've fixed the userdir issue (I think) and added what
> Andreas requested (which was quite self-contained and unlikely to
> break anything else).
Thanks a lot. I see that you also have added the runs test in the Tools >
Nonparametric tests > Runs test menu. That is great. But it still appears
in the Variable menu. Maybe it should be removed from the Variable menu so
it does not appear twice in different menus?
Secondly, I see that the --verbose (print extra output) command can not be
chosen from the Tools > Nonparametric tests > Difference test menu. Could
you please add a check box for this command?
Thirdly, I see that there are no Help buttons for the new Runs test or
Difference test menus. Could these be created by slightly modifying the
runs and difftest command syntax help files?
> Now: please beat on the Windows snapshot and see if it's OK! Thanks.
>
> http://ricardo.ecn.wfu.edu/pub/gretl/gretl_install.exe
I have tried it, and it works without any problems.
Again, thanks for a really great software.
Regards,
Andreas
16 years, 11 months
