matlab to octave code errors

[smed]

Hi, I'm trying to run Matlab code in Octave, and i'm getting the following errors:

warning: function name `fem1d2n' does not agree with function file name `C:\Matlab\fem1d.m'

error: `N' undefined near line 54 column 10
error: evaluating argument list element number 1
error: evaluating for command near line 52, column 1
error: called from `fem1d:InputData' in file `C:\Matlab\fem1d.m'
error: called from `fem1d' in file `C:\Matlab\fem1d.m'

here is the .m file i'm trying to run, thank you for your help:
fem1d.m

-Smed

[Ivan Sutoris]

On Tue, Jun 9, 2009 at 2:55 AM, smed<spontarelliam@...> wrote:

>
> Hi, I'm trying to run Matlab code in Octave, and i'm getting the following
> errors:
>
> warning: function name `fem1d2n' does not agree with function file name
> `C:\Matlab\fem1d.m'
>
> error: `N' undefined near line 54 column 10
> error: evaluating argument list element number 1
> error: evaluating for command near line 52, column 1
> error: called from `fem1d:InputData' in file `C:\Matlab\fem1d.m'
> error: called from `fem1d' in file `C:\Matlab\fem1d.m'
>
> here is the .m file i'm trying to run, thank you for your help:
> http://www.nabble.com/file/p23934669/fem1d.m fem1d.m
>
> -Smed

Hi

If it works in Matlab, then probably the "deal" function works
differently in Octave. You can try to rewrite line 54 in two separate
lines (though I didn't test it, as your program requires additional
input files):

% [N, X(N,:)]=deal(TMP(1),TMP(2:1+NDIM));
N = TMP(1);
X(N,:) = TMP(2:1+NDIM);

Regards
Ivan Sutoris
_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[Jaroslav Hajek-2]

On Tue, Jun 9, 2009 at 2:55 AM, smed<spontarelliam@...> wrote:

>
> Hi, I'm trying to run Matlab code in Octave, and i'm getting the following
> errors:
>
> warning: function name `fem1d2n' does not agree with function file name
> `C:\Matlab\fem1d.m'
>
> error: `N' undefined near line 54 column 10
> error: evaluating argument list element number 1
> error: evaluating for command near line 52, column 1
> error: called from `fem1d:InputData' in file `C:\Matlab\fem1d.m'
> error: called from `fem1d' in file `C:\Matlab\fem1d.m'
>
> here is the .m file i'm trying to run, thank you for your help:
> http://www.nabble.com/file/p23934669/fem1d.m fem1d.m
>
> -Smed
>
>

Could you provide also sample input files so that we can actually try the code?



--
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz
_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[smed]

Ivan Sutoris wrote:

Hi

If it works in Matlab, then probably the "deal" function works
differently in Octave. You can try to rewrite line 54 in two separate
lines (though I didn't test it, as your program requires additional
input files):

% [N, X(N,:)]=deal(TMP(1),TMP(2:1+NDIM));
N = TMP(1);
X(N,:) = TMP(2:1+NDIM);

Regards
Ivan Sutoris
_______________________________________________
Help-octave mailing list
[e-mail]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

Thank you for your response. I tried your suggestion, but i'm still getting errors.

warning: function name `fem1d2n' does not agree with function file name `C:\Matlab\fem1d.m'

error: invalid matrix index = 1
error: evaluating argument list element number 1
error: evaluating for command near line 79, column 1
error: called from `fem1d:InputData' in file `C:\Matlab\fem1d.m'
error: called from `fem1d' in file `C:\Matlab\fem1d.m'

FEM1D.INP

Here is the input file I'm using, and yes it works in matlab. Thanks again.

[John W. Eaton-3]

On  9-Jun-2009, smed wrote:

| Ivan Sutoris wrote:
| >
| > Hi
| >
| > If it works in Matlab, then probably the "deal" function works
| > differently in Octave. You can try to rewrite line 54 in two separate
| > lines (though I didn't test it, as your program requires additional
| > input files):
| >
| > % [N, X(N,:)]=deal(TMP(1),TMP(2:1+NDIM));
| > N = TMP(1);
| > X(N,:) = TMP(2:1+NDIM);

It's not a problem with deal, but with the way multi-assignments are
handled.  Apparently Matlab is creating the variable N before
evaluating the second element in the assignment and Octave is not.
This problem should probably be fixed, but as Ivan notes, you can
easily work around it yourself by using a temporary variable.

| Thank you for your response. I tried your suggestion, but i'm still getting
| errors.
|
| warning: function name `fem1d2n' does not agree with function file name
| `C:\Matlab\fem1d.m'
|
| error: invalid matrix index = 1
| error: evaluating argument list element number 1
| error: evaluating for command near line 79, column 1
| error: called from `fem1d:InputData' in file `C:\Matlab\fem1d.m'
| error: called from `fem1d' in file `C:\Matlab\fem1d.m'
|
| http://www.nabble.com/file/p23941353/FEM1D.INP FEM1D.INP
|
| Here is the input file I'm using, and yes it works in matlab. Thanks again.

OK, this problem seems to be happening here:

  DUMMY = fgets(LINP)
  F = zeros(NN,1);
  for I=1:NL
     TMP = str2num(fgets(LINP));
     [N,tmp]=deal(TMP(1),TMP(2));
     F(N) = tmp;
  end

when indexing TMP in the call to deal.  If I look at the value of TMP,
it is [], because fgets(LINP) is returning "DOF#  Load".  Are you sure
FEM1D.INP is a valid input file for your program?  It seems that ND is
1, but there are 4 lines of displacements.  Is that correct?

jwe

_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[Przemek Klosowski]

First, you get a warning because Matlab language is biased towards a
development style where procedures are written one at a time into
individual *.m files named after the function defined in the file
(i.e. file xyz.m contains  function x=xyz(a,b) .... endfunction).
You can break this assumption but then Octave and Matlab will have harder
time finding the code, and some nice things won't work, like for instance
reloading the file if you edit the procedure while running the interpreter.
As you can see, this is a warning, so you can ignore it if you prefer.

Unfortunately, your code seems to contain errors: specifically, I think
that it uses uninitialized variable N. From the context, I would guess
that the authors confused I and N:

%----- Coordinates -----
DUMMY = fgets(LINP);
for I=1:NN
   TMP = str2num(fgets(LINP));
   [N, X(N,:)]=deal(TMP(1),TMP(2:1+NDIM));
end


_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[John W. Eaton-3]

On  9-Jun-2009, John W. Eaton wrote:

| On  9-Jun-2009, smed wrote:
|
| | Ivan Sutoris wrote:
| | >
| | > Hi
| | >
| | > If it works in Matlab, then probably the "deal" function works
| | > differently in Octave. You can try to rewrite line 54 in two separate
| | > lines (though I didn't test it, as your program requires additional
| | > input files):
| | >
| | > % [N, X(N,:)]=deal(TMP(1),TMP(2:1+NDIM));
| | > N = TMP(1);
| | > X(N,:) = TMP(2:1+NDIM);
|
| It's not a problem with deal, but with the way multi-assignments are
| handled.  Apparently Matlab is creating the variable N before
| evaluating the second element in the assignment and Octave is not.
| This problem should probably be fixed, but as Ivan notes, you can
| easily work around it yourself by using a temporary variable.

I looked at fixing this problem, but I don't see a simple solution.
I'm starting a new thread on the maintainers list about it.

jwe
_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[smed]

I apologize, you were right about the input file having a problem. This is the correct .inp file that works in MatLab, however, it still gives me trouble in octave. After applying your fix to line 54, I now get:

error: invalid vector index = 0

error: evaluating binary operator `-' near line 144, column 16

error: evaluating assignment expression near line 144, column 8

error: evaluating for command near line 140, column 1

-Smed

On Tue, Jun 9, 2009 at 14:23, John W. Eaton <jwe@...> wrote:

On  9-Jun-2009, John W. Eaton wrote:

| On  9-Jun-2009, smed wrote:
|
| | Ivan Sutoris wrote:
| | >
| | > Hi
| | >
| | > If it works in Matlab, then probably the "deal" function works
| | > differently in Octave. You can try to rewrite line 54 in two separate
| | > lines (though I didn't test it, as your program requires additional
| | > input files):
| | >
| | > % [N, X(N,:)]=deal(TMP(1),TMP(2:1+NDIM));
| | > N = TMP(1);
| | > X(N,:) = TMP(2:1+NDIM);
|
| It's not a problem with deal, but with the way multi-assignments are
| handled.  Apparently Matlab is creating the variable N before
| evaluating the second element in the assignment and Octave is not.
| This problem should probably be fixed, but as Ivan notes, you can
| easily work around it yourself by using a temporary variable.

I looked at fixing this problem, but I don't see a simple solution.
I'm starting a new thread on the maintainers list about it.

jwe

_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[John W. Eaton-3]

On  9-Jun-2009, Adam Spontarelli wrote:

| I apologize, you were right about the input file having a problem. This is
| the correct .inp file that works in MatLab, however, it still gives me
| trouble in octave. After applying your fix to line 54, I now get:
|
| error: invalid vector index = 0
|
| error: evaluating binary operator `-' near line 144, column 16
|
| error: evaluating assignment expression near line 144, column 8
|
| error: evaluating for command near line 140, column 1

There are other places where you need a similar change.  Also, some
might not be obvious, or detected in the same as the first.  For
example, once N is defined on line 54, the expression on lines 63 and
64:

   [N,NOC(N,:), MAT(N,:), AREA(N,:), DT(N,:)] = ...
      deal(TMP(1),TMP(2:1+NEN), TMP(2+NEN), TMP(3+NEN), TMP(4+NEN));

will probably not do what you expect.

In any case, it seems to me that all the uses of deal in your file are
somewhat questionable.  I see no good reason to write the above with a
call to deal.  Instead, I would recommend writing this expression as
the series of assignments

  N = TMP(1);
  NOC(N,:) = TMP(2:1+NEN);
  MAT(N,:) = TMP(2+NEN);
  AREA(N,:) = TMP(3+NEN);
  DT(N,:) = TMP(4+NEN);

After I made changes like this, your example worked up to the point of
calling the function "bansol", which you did not provide and which is
not part of Octave (or Matlab, as far as I can tell).

jwe

_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[smed]

Sorry I did not include bandsol.m . I obviously did not write this code, it was provided in a textbook that i'm trying to use for a project. I ran the program with your changes and it worked, but the answers came out as "NaN" which is junk. Are you certain that the change you made didn't affect the logic?

Thank you.

On Tue, Jun 9, 2009 at 15:01, John W. Eaton <jwe@...> wrote:

On  9-Jun-2009, Adam Spontarelli wrote:

| I apologize, you were right about the input file having a problem. This is
| the correct .inp file that works in MatLab, however, it still gives me
| trouble in octave. After applying your fix to line 54, I now get:
|
| error: invalid vector index = 0
|
| error: evaluating binary operator `-' near line 144, column 16
|
| error: evaluating assignment expression near line 144, column 8
|
| error: evaluating for command near line 140, column 1

There are other places where you need a similar change.  Also, some
might not be obvious, or detected in the same as the first.  For
example, once N is defined on line 54, the expression on lines 63 and
64:

  [N,NOC(N,:), MAT(N,:), AREA(N,:), DT(N,:)] = ...

     deal(TMP(1),TMP(2:1+NEN), TMP(2+NEN), TMP(3+NEN), TMP(4+NEN));


will probably not do what you expect.

In any case, it seems to me that all the uses of deal in your file are
somewhat questionable.  I see no good reason to write the above with a
call to deal.  Instead, I would recommend writing this expression as
the series of assignments

 N = TMP(1);
 NOC(N,:) = TMP(2:1+NEN);
 MAT(N,:) = TMP(2+NEN);
 AREA(N,:) = TMP(3+NEN);
 DT(N,:) = TMP(4+NEN);

After I made changes like this, your example worked up to the point of
calling the function "bansol", which you did not provide and which is
not part of Octave (or Matlab, as far as I can tell).

jwe

_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave