5:42 p.m.
Hi,
shouldn't this work? Getting an unmatched data types or something error.
thanks,
sven
<hansl>
nulldata 2
series a = 10 # ID 2 (1 is index)
series b = 11 # ID 3
eval max(deflist(1,2,3))
</hansl>
6:14 p.m.
On Wed, 14 Mar 2018, Sven Schreiber wrote:
Hi,
shouldn't this work? Getting an unmatched data types or something error.
thanks,
sven
<hansl>
nulldata 2
series a = 10 # ID 2 (1 is index)
series b = 11 # ID 3
eval max(deflist(1,2,3))
</hansl>
Yes, should work according to the doc. This usage is an exception to
the rule (recently enforced properly) that series can never be
referred to by ID number when they appear as function arguments. But
since the usage is stated to be OK I have re-enabled it.
Allin
6:22 p.m.
Am 14.03.2018 um 19:14 schrieb Allin Cottrell:
On Wed, 14 Mar 2018, Sven Schreiber wrote:
> eval max(deflist(1,2,3))
> </hansl>
Yes, should work according to the doc. This usage is an exception to the
rule (recently enforced properly) that series can never be referred to
by ID number when they appear as function arguments. But since the usage
is stated to be OK I have re-enabled it.
BTW, this works:
eval max(20 * ones(3,1))
while this throws an error about wrong arg type:
eval max(20 * ones(3,2))
I tend to think that either both should work or both should fail. In the
former case the docs would need to reflect this.
And to be clear in the first case, there are no series IDs 20, so it
could not be possibly interpreted as such.
(There have been several times in the past when I remarked on the list
that max seems to be working for matrix input, but never got much of a
reply.)
thanks,
sven
6:33 p.m.
On Wed, 14 Mar 2018, Sven Schreiber wrote:
Am 14.03.2018 um 19:14 schrieb Allin Cottrell:
> On Wed, 14 Mar 2018, Sven Schreiber wrote:
>> eval max(deflist(1,2,3))
>> </hansl>
>
> Yes, should work according to the doc. This usage is an exception to the
> rule (recently enforced properly) that series can never be referred to by
> ID number when they appear as function arguments. But since the usage is
> stated to be OK I have re-enabled it.
BTW, this works:
eval max(20 * ones(3,1))
while this throws an error about wrong arg type:
eval max(20 * ones(3,2))
I tend to think that either both should work or both should fail. In the
former case the docs would need to reflect this.
I agree. This is a legacy thing -- accepting a column vector as if it
were a series. But since maxc and maxr are available for matrices
that's a bad idea. If we want max to work for matrices we should make
it return the global maximum value, and document that. I'm not sure I
want to make that change right now, however.
And to be clear in the first case, there are no series IDs 20, so it
could
not be possibly interpreted as such.
gretl wouldn't be tempted by that interpretation in any case: it's
only deflist(), not max(), that accepts numerical series IDs.
Allin
6:37 p.m.
Am 14.03.2018 um 19:33 schrieb Allin Cottrell:
On Wed, 14 Mar 2018, Sven Schreiber wrote:
I agree. This is a legacy thing -- accepting a column vector as if it
were a series. But since maxc and maxr are available for matrices that's
a bad idea. If we want max to work for matrices we should make it return
the global maximum value, and document that. I'm not sure I
want to make that change right now, however.
Right. Should we put it on the bug tracker, or are you going to change
it right after the release?
thanks,
sven
6:45 p.m.
On Wed, 14 Mar 2018, Sven Schreiber wrote:
Am 14.03.2018 um 19:33 schrieb Allin Cottrell:
> On Wed, 14 Mar 2018, Sven Schreiber wrote:
> I agree. This is a legacy thing -- accepting a column vector as if it were
> a series. But since maxc and maxr are available for matrices that's a bad
> idea. If we want max to work for matrices we should make it return the
> global maximum value, and document that. I'm not sure I
> want to make that change right now, however.
Right. Should we put it on the bug tracker, or are you going to change it
right after the release?
I'm thinking of right after the release.
Allin
9:49 a.m.
> eval max(deflist(1,2,3))
> </hansl>
Yes, should work according to the doc. This usage is an exception to the
rule (recently enforced properly) that series can never be referred to
by ID number when they appear as function arguments. But since the usage
is stated to be OK I have re-enabled it.
BTW, this works:
eval max(20 * ones(3,1))
while this throws an error about wrong arg type:
eval max(20 * ones(3,2))
It seems there are somewhat different rules work:
eval max(deflist(1,2,3)) gives error since
eval deflist(1,2,3) gives error: deflist requires
declaring named list at the left-hand side
In eval max(20 * ones(3,2)) '20' seems to be
irrelevant: eval max(ones(3,2)) will give an error
Statistical functions except for sum() (max, min, mean, median, etc.)
all works only with vector; one should evaluate max(vec(ones(3,2)) etc
Oleh
14 березня 2018, 20:22:11, від "Sven Schreiber" <svetosch(a)gmx.net>:
Am 14.03.2018 um 19:14 schrieb Allin Cottrell:
> On Wed, 14 Mar 2018, Sven Schreiber wrote:
>> eval max(deflist(1,2,3))
>> </hansl>
>
> Yes, should work according to the doc. This usage is an exception to the
> rule (recently enforced properly) that series can never be referred to
> by ID number when they appear as function arguments. But since the usage
> is stated to be OK I have re-enabled it.
BTW, this works:
eval max(20 * ones(3,1))
while this throws an error about wrong arg type:
eval max(20 * ones(3,2))
I tend to think that either both should work or both should fail. In the
former case the docs would need to reflect this.
And to be clear in the first case, there are no series IDs 20, so it
could not be possibly interpreted as such.
(There have been several times in the past when I remarked on the list
that max seems to be working for matrix input, but never got much of a
reply.)
thanks,
sven
_______________________________________________
Gretl-devel mailing list
Gretl-devel@lists.wfu.eduhttp://lists.wfu.edu/mailman/listinfo/gretl-devel
10:30 a.m.
Am 15.03.2018 um 10:49 schrieb oleg_komashko(a)ukr.net:
>> eval max(deflist(1,2,3))
>> </hansl>
>
[Allin:]
> Yes, should work according to the doc. This usage is an exception
to the
> rule (recently enforced properly) that series can never be referred to
> by ID number when they appear as function arguments. But since the usage
> is stated to be OK I have re-enabled it.
[Oleh:]
It seems there are somewhat different rules work:
eval max(deflist(1,2,3)) gives error since
eval deflist(1,2,3) gives error: deflist requires
declaring named list at the left-hand side
No, it was a bug and should work in the future, see above.
In eval max(20 * ones(3,2)) '20' seems to be
irrelevant:
Yes, I just wanted to make sure those numbers could not be mistaken for
series IDs (in a workfile with just 2 or 3 series).
eval max(ones(3,2)) will give an error
Statistical functions except for sum() (max, min, mean, median,
etc.)
all works only with vector; one should evaluate max(vec(ones(3,2)) etc
If you mean 'vector' instead of a matrix with more than one row/col,
that's what Allin referred to in his earlier answer. (legacy thing)
(Because those functions do not _only_ work with a vector [or series],
they also work on lists, but with a slightly different meaning, see the
doc.)
cheers,
sven
4:35 p.m.
If you mean 'vector' instead of a matrix with more than one
row/col,
that's what Allin referred to in his earlier answer. (legacy thing)
max(vec(a_matrix)) works better them imaxc(imaxr(a_matrix))
I have run it with $stopwatch and huge matrices
Oleh
15 березня 2018, 12:30:29, від "Sven Schreiber" <svetosch(a)gmx.net>:
Am 15.03.2018 um 10:49 schrieb > oleg_komashko(a)ukr.net:
>>> eval max(deflist(1,2,3))
>>> </hansl>
>>
[Allin:]
>> Yes, should work according to the doc. This usage is an exception to the
>> rule (recently enforced properly) that series can never be referred to
>> by ID number when they appear as function arguments. But since the usage
>> is stated to be OK I have re-enabled it.
[Oleh:]
> It seems there are somewhat different rules work:
>
> eval max(deflist(1,2,3)) gives error since
>
> eval deflist(1,2,3) gives error: deflist requires
> declaring named list at the left-hand side
No, it was a bug and should work in the future, see above.
> In eval max(20 * ones(3,2)) '20' seems to be
> irrelevant:
Yes, I just wanted to make sure those numbers could not be mistaken for
series IDs (in a workfile with just 2 or 3 series).
eval max(ones(3,2)) will give an error
> Statistical functions except for sum() (max, min, mean, median, etc.)
> all works only with vector; one should evaluate max(vec(ones(3,2)) etc
If you mean 'vector' instead of a matrix with more than one row/col,
that's what Allin referred to in his earlier answer. (legacy thing)
(Because those functions do not _only_ work with a vector [or series],
they also work on lists, but with a slightly different meaning, see the
doc.)
cheers,
sven
_______________________________________________
Gretl-devel mailing list
Gretl-devel@lists.wfu.eduhttp://lists.wfu.edu/mailman/listinfo/gretl-devel
2787
days inactive
2788
days old
8 comments
3 participants
participants (3)
-
Allin Cottrell -
oleg_komashko＠ukr.net -
Sven Schreiber