[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: why two indexes in 1d cell array ? (octave-3.4.2) ... how about more
From: |
Ben Abbott |
Subject: |
Re: why two indexes in 1d cell array ? (octave-3.4.2) ... how about more dimensions ? |
Date: |
Wed, 10 Aug 2011 19:06:36 -0400 |
On Aug 10, 2011, at 6:22 PM, Ben Abbott wrote:
> On Aug 10, 2011, at 5:33 PM, Sergei Steshenko wrote:
>
>> --- On Wed, 8/10/11, Ben Abbott <address@hidden> wrote:
>>
>>> From: Ben Abbott <address@hidden>
>>> Subject: Re: why two indexes in 1d cell array ? (octave-3.4.2) ... how
>>> about more dimensions ?
>>> To: "Sergei Steshenko" <address@hidden>
>>> Cc: address@hidden
>>> Date: Wednesday, August 10, 2011, 2:06 PM
>>> On Aug 10, 2011, at 3:18 PM, Sergei
>>> Steshenko wrote:
>>>
>>>> c =
>>>> {
>>>> [1,1] = a string
>>>> [1,2] =
>>>> "
>>>>
>>>> I see _two_ indexes, i.e. "[1,1]...", "[1,2]", though
>>> to me it looks like
>>>> created a _1d_ (just _one_ dimension) cell array;
>>>
>>> Matlab originated from Fortran code. Thus, arrays were/are
>>> column-major.
>>>
>>> http://en.wikipedia.org/wiki/Row-major_order#Column-major_order
>>>
>>> When the source code switched to C, Matlab maintained the
>>> default column-major storage order for arrays. This means
>>> that A = 1:5 produces a single row with 5 columns.
>>>
>>> However, since C is row-major ...
>>>
>>> http://en.wikipedia.org/wiki/Row-major_order#Row-major_order
>>>
>>> ... a 1D array requires two indices. The fist specifies the
>>> row [1], and the second the column.
>>>
>>> For compatibility with Matlab, Octave adopted the same
>>> approach.
>>>
>>>> Is this all expected and documented behavior ?
>>>
>>> Yes this is expected and deliberate. If you search the
>>> manual. you'll find a few instances where the effect of
>>> "column-major" ordering is mentioned.
>>>
>>> Ben
>>
>> Let's again have a look at documentation:
>>
>> http://www.gnu.org/software/octave/doc/interpreter/Basic-Usage-of-Cell-Arrays.html#Basic-Usage-of-Cell-Arrays
>> .
>>
>> I am going to repeat actions described in the documentation:
>>
>> 1a) documentation:
>> c = {"a string", rand(2, 2)};
>>
>> 1b) me:
>> "
>> c = {"a string", rand(2, 2)}
>> c =
>> {
>> [1,1] = a string
>> [1,2] =
>>
>> 0.64730 0.79055
>> 0.40904 0.54691
>>
>> }
>> "
>> - 'octave' prints differently, but I didn't put ';' at the end of my input,
>> so IMO in this case 'octave' behaves as expected;
>>
>>
>> 2a) documentation:
>> "
>> c{1:2}
>> ⇒ ans =
>>
>> (,
>> [1] = a string
>> [2] =
>>
>> 0.593993 0.627732
>> 0.377037 0.033643
>>
>> ,)
>> "
>> 2b) me:
>> "
>> c{1:2}
>> ans = a string
>> ans =
>>
>> 0.64730 0.79055
>> 0.40904 0.54691
>> ".
>>
>> In '2b)' _no_ indexes.
>>
>> Why the discrepancy between actual behavior and the documentation ?
>>
>> Thanks,
>> Sergei.
>
> You've found an error in the documentation.
>
> If someone doesn't get to is before me, I'll submit a fix for it later today.
>
> Ben
Done.
http://hg.savannah.gnu.org/hgweb/octave/rev/1acb639204d0
Ben