4:19 p.m.
Hi,
I have a monthly time series of the Dow Jones Industrial Index beginning
1928:10. I'd like to create a new variable that is just the month
number: 10, 11, 12, 1, 2, 3,.... How can I do this? I thought there
would be a month function in Gretl, but I don't see one. This question,
of course, extends to the year (1928, 1928, 1929, 1929,...) and quarter
if I have quarterly data.
Thanks,
Walt
________________________
Walter R. Paczkowski, Ph.D.
Data Analytics Corp.
44 Hamilton Lane
Plainsboro, NJ 08536
________________________
(V) 609-936-8999
(F) 609-936-3733
walt(a)dataanalyticscorp.com
www.dataanalyticscorp.com
_____________________________________________________
8:40 p.m.
On Sun, 16 Sep 2012, Data Analytics Corp. wrote:
Hi,
I have a monthly time series of the Dow Jones Industrial Index beginning
1928:10. I'd like to create a new variable that is just the month
number: 10, 11, 12, 1, 2, 3,.... How can I do this? I thought there
would be a month function in Gretl, but I don't see one. This question,
of course, extends to the year (1928, 1928, 1929, 1929,...) and quarter
if I have quarterly data.
Does this help?
<hansl>
open bjg.gdt
series year = 1949 + floor((t-1)/12)
series month = 1 + ((t-1)%12)
</hansl>
--------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
--------------------------------------------------
5:45 a.m.
Very compact, as always. But it wouldn't work for a dataset with weekly or daily
frequency, would it?
Jan Annaert
UNIVERSITEITANTWERPEN | Faculty of Applied Economics (TEW) | Dept. Accounting &
Finance
Room S.B.335 | Prinsstraat 13 | B-2000 Antwerp | Belgium
Phone +32 32654163 |Fax +32 32654064
http://www.ua.ac.be/jan.annaert
http://ssrn.com/author=143473
From: "Riccardo Lucchetti (Jack)"
<r.lucchetti@univpm.it<mailto:r.lucchetti@univpm.it>>
Reply-To: "r.lucchetti@univpm.it<mailto:r.lucchetti@univpm.it>"
<r.lucchetti@univpm.it<mailto:r.lucchetti@univpm.it>>,
"gretl-users@lists.wfu.edu<mailto:gretl-users@lists.wfu.edu>"
<gretl-users@lists.wfu.edu<mailto:gretl-users@lists.wfu.edu>>
Date: zondag 16 september 2012 22:40
To: "walt@dataanalyticscorp.com<mailto:walt@dataanalyticscorp.com>"
<walt@dataanalyticscorp.com<mailto:walt@dataanalyticscorp.com>>,
"gretl-users@lists.wfu.edu<mailto:gretl-users@lists.wfu.edu>"
<gretl-users@lists.wfu.edu<mailto:gretl-users@lists.wfu.edu>>
Subject: Re: [Gretl-users] extracting month number
On Sun, 16 Sep 2012, Data Analytics Corp. wrote:
Hi,
I have a monthly time series of the Dow Jones Industrial Index beginning
1928:10. I'd like to create a new variable that is just the month
number: 10, 11, 12, 1, 2, 3,.... How can I do this? I thought there
would be a month function in Gretl, but I don't see one. This question,
of course, extends to the year (1928, 1928, 1929, 1929,...) and quarter
if I have quarterly data.
Does this help?
<hansl>
open bjg.gdt
series year = 1949 + floor((t-1)/12)
series month = 1 + ((t-1)%12)
</hansl>
--------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti@univpm.it<mailto:r.lucchetti@univpm.it>
http://www2.econ.univpm.it/servizi/hpp/lucchetti
--------------------------------------------------
6:22 a.m.
On Mon, 17 Sep 2012, Annaert Jan wrote:
Very compact, as always. But it wouldn't work for a dataset with
weekly
or daily frequency, would it?
For daily data, I can't see why not. For weekly data, the problem is that
365 is not a multiple of 7.
--------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
--------------------------------------------------
9:29 a.m.
On 16/09/12 22:40, Riccardo (Jack) Lucchetti wrote:
...
Does this help?
<hansl>
open bjg.gdt
series year = 1949 + floor((t-1)/12)
series month = 1 + ((t-1)%12)
</hansl>
--------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Economia
A different possibility is to play with strings. We have the command
'obslabel' for this.
For example, if the observations are quarterly, this code writes the
year of the 't' observation into the string 'iyear' and the quarter into
the string 'iquarter': (where t=1, 2, 3, ... is the 1-based index of the
observation: )
<hansl>
string s = obslabel(t)
sscanf s, "%d:%d", tyear, tquarter
</hansl>
If you need it, you may then put the strings in scalars in this way:
scalar ty=@tyear
scalar tq=@tquarter
--
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
9:52 a.m.
On 17/09/12 11:29, Ignacio Diaz-Emparanza wrote:
A different possibility is to play with strings. We have the command
'obslabel' for this. For example, if the observations are quarterly,
this code writes the year of the 't' observation into the string
'iyear' and the quarter into the string 'iquarter':
Sorry, in the script the strings were defined as "tyear" and
"tquarter"
instead of "iyear" and "iquarter".
--
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
12:32 p.m.
On Mon, 17 Sep 2012, Ignacio Diaz-Emparanza wrote:
A different possibility is to play with strings. We have the command
'obslabel' for this.
For example, if the observations are quarterly, this code writes the
year of the 't' observation into the string 'iyear' and the quarter into
the string 'iquarter': (where t=1, 2, 3, ... is the 1-based index of the
observation: )
<hansl>
string s = obslabel(t)
sscanf s, "%d:%d", tyear, tquarter
</hansl>
If you need it, you may then put the strings in scalars in this way:
scalar ty=@tyear
scalar tq=@tquarter
Hmm, the "%d" conversion already produces a scalar.
If series results are wanted, here's a variant on Ignacio's
suggestion. It's bulkier than Jack's suggestion, but it
doesn't depend on knowing the starting year and quarter in
advance:
<hansl>
open data9-7
series quarter year
scalar tyear tquart
string s
loop t=1..$nobs -q
s = obslabel(t)
sscanf s, "%d:%d", tyear, tquart
year[t] = tyear
quarter[t] = tquart
endloop
print year quarter --byobs
</hansl>
Allin Cottrell
10 p.m.
On Mon, 17 Sep 2012, Allin Cottrell wrote:
If series results are wanted, here's a variant on Ignacio's
suggestion. It's bulkier than Jack's suggestion, but it doesn't
depend on knowing the starting year and quarter in advance:
<hansl>
open data9-7
series quarter year
scalar tyear tquart
string s
loop t=1..$nobs -q
s = obslabel(t)
sscanf s, "%d:%d", tyear, tquart
year[t] = tyear
quarter[t] = tquart
endloop
print year quarter --byobs
</hansl>
Thinking about this, it seemed lame that you had to use temporary
scalar variables to accept the result of scalar conversions from
sscanf -- rather than assigning the results directly to elements of
series (or matrices). That's now fixed: the following nicer version
will work with current CVS:
<hansl>
open data9-7
series quarter year
string s
loop t=1..$nobs -q
s = obslabel(t)
sscanf s, "%d:%d", year[t], quarter[t]
endloop
</hansl>
This will be in the snapshots for Windows and Mac tomorrow.
Allin Cottrell
4493
days inactive
4494
days old
7 comments
5 participants
participants (5)
-
Allin Cottrell -
Annaert Jan -
Data Analytics Corp. -
Ignacio Diaz-Emparanza -
Riccardo (Jack) Lucchetti