[Octave-bug-tracker] [bug #52086] t=1:5; t(zeros(0,1)+[1:2]) returns 0×2 empty double matrix in Matlab

[Alois Schlögl]

URL:
  <http://savannah.gnu.org/bugs/?52086>

                 Summary: t=1:5; t(zeros(0,1)+[1:2]) returns 0×2 empty double
matrix in Matlab 
                 Project: GNU Octave
            Submitted by: schloegl
            Submitted on: Fri 22 Sep 2017 06:37:02 AM UTC
                Category: Octave Function
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Matlab Compatibility
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: 4.2.1
        Operating System: GNU/Linux

    _______________________________________________________

Details:

Recently, I tried to convert some Matlab code to Octave. In Matlab
9.1.0.441655 (R2016b), this expression returned 

   t=1:5; t(zeros(0,1)+[1:2]) 
   0×2 empty double matrix

 
In Octave 4.2.1 the expression fails with this error

   error: operator +: nonconformant arguments (op1 is 0x1, op2 is 1x2)


Matlab R2013a fails with this error

   Error using +
   Matrix dimensions must agree.


If it is a 0x0 or 1x1, like this

   t(zeros(1,0)+[1:2])
   t(zeros(0,0)+[1:2])

   Error using +
   Matrix dimensions must agree.

the expression fails in Matlab R2016b as well.

In the present case, the empty matrix zeros(0,1) was the result of "find(..)",
and under special conditions it can support writing leaner code. One could
argue whether this needs fixing, but Octave should be aware of this.


Best,
   Alois 






    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?52086>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/