11:49 a.m.
Hi there,
why are those not equivalent? I.e., the foreach (which is more concise) does not work as
(I) expected. Is there a tweak to make it work?
<hansl>
open denmark.gdt -q
list ls = 1 4
matrices M = null
loop foreach l ls
matrix temp = l
colnames(temp,varname(l))
M+=temp
endloop
matrices M2 = null
loop for s=1..nelem(ls)
matrix temp = ls[s]
colnames(temp,varname(ls[s]))
M2 += temp
endloop
?M
?M2
</hansl>
Thanks
Frederik
11:54 a.m.
Am 15.12.2017 um 17:49 schrieb Schaff, Frederik:
why are those not equivalent? I.e., the foreach (which is more
concise)
does not work as (I) expected.
As always, please define "as expected". And what are you getting instead?
loop for s=1..nelem(ls)
BTW, this syntax is not optimal (even though it's tolerated) -- for the
simplest index loop omit the "for": loop s=1..nelem(ls) will do.
cheers,
sven
1:56 p.m.
Hi there, I'd expect both options to be equivalent (add a matrix with the data from
the series to the array of matrices). However, the forach loop only adds the first
data-point to the matrix.
As always, please define "as expected". And what are you
getting instead?
<hansl>
open denmark.gdt -q
list ls = 1 4
matrices M = null
loop foreach l ls
matrix temp = l
colnames(temp,varname(l))
M+=temp
endloop
matrices M2 = null
loop s=1..nelem(ls)
matrix temp = ls[s]
colnames(temp,varname(ls[s]))
M2 += temp
endloop
?M
?M2
</hansl>
produces:
<output>
? M
Array of matrices, length 2
[1] 1 x 1
[2] 1 x 1
? M2
Array of matrices, length 2
[1] 55 x 1
[2] 55 x 1
<\output>
> -----Ursprüngliche Nachricht-----
> Von: gretl-users-bounces(a)lists.wfu.edu [mailto:gretl-users-
> bounces(a)lists.wfu.edu] Im Auftrag von Sven Schreiber
> Gesendet: Freitag, 15. Dezember 2017 17:55
> An: gretl-users(a)lists.wfu.edu
> Betreff: Re: [Gretl-users] foreach vs for with list
>
> Am 15.12.2017 um 17:49 schrieb Schaff, Frederik:
>
> > why are those not equivalent? I.e., the foreach (which is more
> > concise) does not work as (I) expected.
>
As always, please define "as expected". And what are you
getting instead?
>
> > loop for s=1..nelem(ls)
>
> BTW, this syntax is not optimal (even though it's tolerated) -- for the simplest
> index loop omit the "for": loop s=1..nelem(ls) will do.
>
> cheers,
> sven
> _______________________________________________
> Gretl-users mailing list
> Gretl-users(a)lists.wfu.edu
> http://lists.wfu.edu/mailman/listinfo/gretl-users
2:33 p.m.
On Fri, 15 Dec 2017, Schaff, Frederik wrote:
Hi there, I'd expect both options to be equivalent (add a matrix
with the data from the series to the array of matrices). However,
the forach loop only adds the first data-point to the matrix.
<hansl>
open denmark.gdt -q
list ls = 1 4
matrices M = null
loop foreach l ls
matrix temp = l
colnames(temp,varname(l))
M+=temp
endloop
matrices M2 = null
loop s=1..nelem(ls)
matrix temp = ls[s]
colnames(temp,varname(ls[s]))
M2 += temp
endloop
In your first loop, 'l' is a just scalar index, so
matrix temp = l
gives you {1} on the first iteration and {2} on the second. If you
want to access the series at position l in the list you need to say
matrix temp = $l
Then you'll get the same output as from your second loop.
Allin
2:40 p.m.
On Fri, 15 Dec 2017, Allin Cottrell wrote:
On Fri, 15 Dec 2017, Schaff, Frederik wrote:
> Hi there, I'd expect both options to be equivalent (add a matrix with the
> data from the series to the array of matrices). However, the forach loop
> only adds the first data-point to the matrix.
>
> <hansl>
> open denmark.gdt -q
> list ls = 1 4
> matrices M = null
> loop foreach l ls
> matrix temp = l
> colnames(temp,varname(l))
> M+=temp
> endloop
>
> matrices M2 = null
> loop s=1..nelem(ls)
> matrix temp = ls[s]
> colnames(temp,varname(ls[s]))
> M2 += temp
> endloop
In your first loop, 'l' is a just scalar index, so
matrix temp = l
gives you {1} on the first iteration and {2} on the second. If you want to
access the series at position l in the list you need to say
matrix temp = $l
Then you'll get the same output as from your second loop.
P.S. the varname() call in your first loop also needs fixing: on the
second iteration l = 2 but presumably you want the name of the second
series in the list, namely series 4, so:
colnames(temp,varname(ls[l]))
Allin
9:33 a.m.
Hi Allin,
many thanks and sry for my late help. I'll save that info.
Best
Frederik
-----Ursprüngliche Nachricht-----
Von: gretl-users-bounces(a)lists.wfu.edu [mailto:gretl-users-
bounces(a)lists.wfu.edu] Im Auftrag von Allin Cottrell
Gesendet: Freitag, 15. Dezember 2017 20:40
An: Gretl list <gretl-users(a)lists.wfu.edu>
Betreff: Re: [Gretl-users] foreach vs for with list
On Fri, 15 Dec 2017, Allin Cottrell wrote:
> On Fri, 15 Dec 2017, Schaff, Frederik wrote:
>
>> Hi there, I'd expect both options to be equivalent (add a matrix with
>> the data from the series to the array of matrices). However, the
>> forach loop only adds the first data-point to the matrix.
>>
>> <hansl>
>> open denmark.gdt -q
>> list ls = 1 4
>> matrices M = null
>> loop foreach l ls
>> matrix temp = l
>> colnames(temp,varname(l))
>> M+=temp
>> endloop
>>
>> matrices M2 = null
>> loop s=1..nelem(ls)
>> matrix temp = ls[s]
>> colnames(temp,varname(ls[s]))
>> M2 += temp
>> endloop
>
> In your first loop, 'l' is a just scalar index, so
>
> matrix temp = l
>
> gives you {1} on the first iteration and {2} on the second. If you
> want to access the series at position l in the list you need to say
>
> matrix temp = $l
>
> Then you'll get the same output as from your second loop.
P.S. the varname() call in your first loop also needs fixing: on the second
iteration l = 2 but presumably you want the name of the second series in the
list, namely series 4, so:
colnames(temp,varname(ls[l]))
Allin
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
2484
days inactive
2487
days old
5 comments
3 participants
participants (3)
-
Allin Cottrell -
Schaff, Frederik -
Sven Schreiber