help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: subscripted assignment dimension mismatch (NDArray)


From: Taku J SATO
Subject: Re: subscripted assignment dimension mismatch (NDArray)
Date: Fri, 6 Feb 2004 03:01:27 +0900

hi,

thank you very much for the patch; it works very nicely now.

btw, with octave-2.1.53, we cannot use the third index
in the following case:

-----
octave:1> test=zeros(4,4,1)
test =

  0  0  0  0
  0  0  0  0
  0  0  0  0
  0  0  0  0

octave:2> test(1,1,1)
error: index exceeds N-d array dimensions
-----

to avoid this difficulty, i use:

------
octave:2> test=reshape(zeros(4,4,1),4,4,1)
test =

ans =

  0  0  0  0
  0  0  0  0
  0  0  0  0
  0  0  0  0



octave:3> test(1,1,1)
ans = 0
-----

is this behavior what octave is supposed to do?
(i presently do not have access to matlab, so i am not
sure what matlab would do...)

all the best,
taku


On Feb 5, 2004, at 3:40, John W. Eaton wrote:

On  4-Feb-2004, Taku J SATO <address@hidden> wrote:

| thanks for your quick reply.  perhaps, my example was not good.
| what i want to do is as follows:
|
| -----
| octave:1> test = zeros(4,4,2)
| octave:2> test2 = [1,2]'
| octave:3> test(1:2,1,1) = test2
| error: subscripted assignment dimension mismatch
| error: assignment failed, or no method for `matrix = matrix'
| error: evaluating assignment expression near line 3, column 15
| ------
|
| actually, i am playing with much bigger arrays (such as 5D array
| assigned by 2D matrix.) the above is the simplest case i could
| reproduce the problem.
|
| (i am presently writing a program which produces all the irreducible
| representations of the k-groups of the crystallographic space groups.
| this program seems running on octave-2.1.52 but fails under 2.1.53...)
|
| all the best,
| taku

Please try the following patch.

Thanks,

jwe


2004-02-04  John W. Eaton  <address@hidden>

        * Array.cc (Array<T>::assignN): Simplify dimension check by
        comparing rhs_dims and frozen_len sans trailing singletons.


Index: liboctave/Array.cc
===================================================================
RCS file: /usr/local/cvsroot/octave/liboctave/Array.cc,v
retrieving revision 1.86
diff -c -r1.86 Array.cc
*** liboctave/Array.cc  23 Jan 2004 03:09:59 -0000      1.86
--- liboctave/Array.cc  4 Feb 2004 18:26:11 -0000
***************
*** 2746,2781 ****
      {
        // RHS is matrix or higher dimension.

-       bool dim_ok = true;
-
-       int jj = 0;
-
        // Check that RHS dimensions are the same length as the
!       // corresponding LHS dimensions.
!
!       int rhs_dims_len = rhs_dims.length ();

!       for (int j = 0; j < idx_is_colon.length (); j++)
!       {
!         if (jj < rhs_dims.length () && rhs_dims(jj) == 1)
!           jj++;
!         else if (idx_is_colon(j))
!           {
!             if (jj > rhs_dims_len || rhs_dims(jj) < lhs_dims(j))
!               {
!                 dim_ok = false;
!
!                 break;
!               }
!
!             jj++;
!           }
!       }

!       if (jj != rhs_dims_len)
!       dim_ok = false;

!       if (! dim_ok)
        (*current_liboctave_error_handler)
          ("subscripted assignment dimension mismatch");
        else
--- 2746,2761 ----
      {
        // RHS is matrix or higher dimension.

        // Check that RHS dimensions are the same length as the
!       // corresponding LHS index dimensions.

!       dim_vector t_rhs_dims = rhs_dims;
!       t_rhs_dims.chop_trailing_singletons ();

!       dim_vector t_frozen_len = frozen_len;
!       t_frozen_len.chop_trailing_singletons ();

!       if (t_rhs_dims != t_frozen_len)
        (*current_liboctave_error_handler)
          ("subscripted assignment dimension mismatch");
        else



--
Taku J Sato



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------



reply via email to

[Prev in Thread] Current Thread [Next in Thread]