FAQ: Difference between revisions

The GNU Octave language is quite similar to Matlab so that most programs are easily portable.

==What is Octave-Forge?==

[https://octave.sourceforge.io/ Octave-Forge] is a collection of [[packages]] for GNU Octave, something similar to the Matlab toolboxes.  When talking about the two projects at the same time, GNU Octave is usually referred to as Octave core (or just "core").  Octave-Forge also serves as a test bed for code that may eventually end up in the core, and distributes binaries for systems with a lack of developers tools (mainly Windows).

   John W. Eaton, David Bateman, Søren Hauberg, Rik Wehbring (2017).

     year      = {2017},

Run <code>citation</code> at the Octave prompt for details on how to best cite the Octave version you are running.  Certain Octave packages also have recommended citations in which case use <code>citation package_name</code>.

Besides this wiki, the GNU Octave distribution includes a [http://www.octave.org/doc/interpreter 1000+ page Texinfo manual] ([http://www.octave.org/octave.pdf PDF]).  Access to the complete text of the manual is available via the <code>doc</code> command at the GNU Octave prompt.  If you have problems using this manual, or find that some topic is not adequately explained, indexed, or cross-referenced, please report it on http://bugs.octave.org.

The used version of Octave is available via the <code>ver</code> command and a list of user-visible changes since the last release is available via the <code>news</code> command at the GNU Octave prompt.

* http://www.octave.org/download.html

* The pkg command now accepts a -forge option to pull packages directly from Octave-forge

=Octave-Forge=

== How do install or load all Octave-Forge packages? ==

== How do I automatically load a package at Octave startup? ==

First of all, make sure you understand [http://www.gnu.org/software/octave/doc/interpreter/Script-Files.html the difference between script files and function files]. If you want to execute a function defined in a file, just call the function like any other Octave function: <code>foo(arg1, arg2);</code>

To execute a script from within Octave, just type its name without the .m extension. Thus, if you have a script called <code>foo.m</code>, just type <code>foo</code> from within Octave to execute it. You have to make sure that the script is in your current path. Type <code>path</code> in Octave to see what this path is, and type <code>pwd</code> to print the working directory (where you're currently standing). The current working directory is referred to as "." in the <code>path</code>.

If the script name has characters that are not valid for an Octave identifier, or if you do not want to use addpath to add the script's location to the current path, you can use the <code>run</code> function instead:

  octave> run("Script Name With Spaces.m")

  octave> run("/opt/local/foo.m")

An alternative is to run the script from outside Octave by calling Octave from your operating system shell. Unlike calling the script from inside Octave, this also allows you to pass arguments from the shell into the script, which the script can access using the <code>argv</code> command:

==How do I erase a figure?==  

    octave:3> format short

    octave:4> pi

==How do I call an octave function from C++?==

* Here is an untested code snippet for calling rand([9000,1]), modified from a post by Herber Farnsworth? to help-octave on 2003-05-01:

#include <octave/oct.h>

...

Matrix unis(f_ret(0).matrix_value());

Look at functions like exist, file_in_path.. and the other functions that their descriptions point to.

'''This only works with gnuplot as graphics_toolkit, NOT with fltk. See [https://savannah.gnu.org/bugs/?33180 Bug#33180]'''

  plot(sin(1:100));

One can set that behaviour as default:

  set(0, 'defaultfigurevisible', 'off');

==How do I make Octave use more precision?==

Octave's default numerical type is IEEE 754 doubles, a.k.a. hardware floats. This type has 52 bits of precision or about 16 decimal digits. It's implemented in your computer's hardware, in your CPU, so it's '''fast'''. This type is assumed throughout for Octave's calculations.

You can use a few other built-in types. The int64 type will have 63 bits of precision. One bit is used for the sign, but if you don't want to lose that bit, uint64 can be used instead. These types, however, cannot represent numbers as large as the default double type, and can only represent integers. Furthermore, there is no way to represent integer literals, so if you do

<syntaxhighlight lang="Octave">uint64(18446744073709551610);</syntaxhighlight>

 

the literal "18446744073709551610" first gets converted to a double precision type, so <code>uint64</code>'s additional precision is lost. Instead, initialize the <code>uint64</code> with smaller numbers and perform a computation to get the larger number you want. E.g.,

 

<syntaxhighlight lang="Octave">uint64(999999999999999) * 10000</syntaxhighlight>

would produce value 9999999999999990000, which is close to the maximum possible value for the uint64 type, but can't be at the moment input directly, doing uint64(9999999999999990000), due to the mentioned error of rounding.

Alternatively, one may use arbitrary precision arithmetic, which has as much precision as is practical to hold in your computer's memory. The ''symbolic'' package has a vpa() function for arbitrary precision arithmetic. Note that arbitrary precision arithmetic must be implemented '''in software''' which makes it much slower than hardware floats.

Consider carefully if your problem really needs more precision. Often if you're running out of precision the problem lies fundamentally in your methods being [http://en.wikipedia.org/wiki/Numerical_stability numerically unstable], so more precision will not help you here. If you absolutely must use arbitrary-precision arithmetic, you're at present better off using a CAS instead of Octave. An example of such a CAS is [http://sagemath.org Sage].

You can't. Matlab P-files (files with a .p file extension), also known as P-code, are [https://en.wikipedia.org/wiki/Obfuscation_%28software%29 obfuscated] files than cannot be run outside of Matlab itself. The original source Matlab m-files that were used to generate the P-files should be used in Octave instead.

In addition to consulting Octave's source for the precise details, you can read the Octave manual for a complete high-level description of the algorithm that Octave uses to decide how to solve a particular linear system, e.g. how the backslash operator {{Codeline|A\x}} will be interpreted.  Sections [http://www.octave.org/doc/interpreter/Techniques-Used-for-Linear-Algebra.html Techniques Used for Linear Algebra] and [http://www.octave.org/doc/interpreter/Sparse-Linear-Algebra.html Linear Algebra on Sparse Matrices] from the manual describe this procedure.

You are probably looking for the function ''lookfor''. This function searches the help text of all functions for a specific string and returns a list of functions. Note that by default it will only search the first line of the help text (check ''help lookfor'' at the octave prompt for more). The following example helps to find the function to calculate correlation coefficient in a matrix:

   octave> lookfor correlation

   corr2              Returns the correlation coefficient between I and J.

  cor                Compute correlation.

   corrcoef            Compute correlation.

   spearman            Compute Spearman's rank correlation coefficient RHO for each of the variables sp

  autocor            Return the autocorrelations from lag 0 to H of vector X.

Also, there's a high chance that the function name has a suggestive name, and so you can try auto-completion to get some hints. For the previous example, typing ''corr'' at the octave promp followed by pressing [Tab] twice would suggest the following:

   octave> corr

   corr2     corrcoef

==I am running a script that should produce output during execution but I don't see anything until it has finished==

By default Octave is set to pass its screen output through a pager (usually the default pager is "less") which allows

things such as navigating through the output with arrow keys or searching for text or regular expressions within the output.

The pager only displays the output after it's finished receiving it, so when it is active you'll not be able to see anything until your script has terminated. To change this behavior temporarily or permanently you may want to use one of the options described [http://www.octave.org/doc/interpreter/Paging-Screen-Output.html here].

If you are running an Octave script that includes a plotting command, the script and Octave may terminate immediately. So the plot window does show up, but immediately closes when Octave finishes execution. Alternatively, if using fltk, the plot window needs a readline loop to show up (the time when Octave is sitting around doing nothing waiting for interactive input).

 

 

 

 

You should normally use your distribution's packages. For Debian and Fedora, Octave package '''foo''' will be a deb or rpm called '''octave-foo''', and you should install that instead using apt or yum.

If you really need to build Octave packages from source to install them, you'll need mkoctfile. Most distributions split Octave into several packages. The script mkoctfile is then part of a separate package:

* Debian/Ubuntu: {{Codeline|octave-headers}} or {{Codeline|liboctave-dev}}

* Fedora: {{Codeline|octave-devel}}

==I'm having problem X using the latest Octave version==

Please be more specific. What is the latest version, according to you? If you mean the latest released version, be aware that you may still have an older version due to whatever distribution method you're using. There may be a newer version available that you're not aware of due to the distribution method you're using to get Octave, and in most cases, there is a way to get a newer version via your distribution method (see other wiki pages for [[Octave_for_GNU/Linux|GNU/Linux]], [[Octave_for_MacOS_X|Mac OSX]], and [[Octave_for_Windows|Windows]]).

If you mean the latest Mercurial revision, please specify which one that is. "Latest tip" is not informative, because from the time you wrote "latest tip" to the time someone reads that message, "latest tip" might have changed meaning. Also, you might be standing on a different commit than what "hg tip" says. The tip may be on a different branch, or you might have updated to a different revision different from what "hg tip" says.

Instead, report the output of "hg summary" or "hg id". Also please use hashes instead or in addition to revision numbers. Revision numbers are just a convenience and only make sense in your local repo, and might not coincide with what someone sees on their own repo. Hashes are globally unique across all repos.

If your problem truly persists with the latest version, as indicated [http://www.gnu.org/software/octave/download.html here], then by all means report a bug or ask for help, but don't be surprised if volunteers are less inclined to help you with a problem that only exists in an older version of Octave.

Floating point arithmetic is an approximation '''in binary''' to arithmetic on real or complex numbers. Just like you cannot represent 1/3 exactly in decimal arithmetic (0.333333 is only a rough approximation to 1/3), you cannot represent some fractions like <math>1/10</math> exactly in base 2. In binary, the representation to one tenth is <math>0.0\overline{0011}_b</math> where the bar indicates that it repeats infinitely (like how <math>1/6 = 0.1\overline{6}_d</math> in decimal). Because this infinite repetition cannot be represented exactly with a finite number of digits, rounding errors occur for values that appear to be exact in decimal but are in fact approximations in binary, such as for example how 0.3 - 0.2 - 0.1 is not equal to zero.

 

 

This isn't an Octave bug. It happens with any program that uses [http://en.wikipedia.org/wiki/IEEE_754 IEEE 754 floating point arithmetic]. To be fair, IEEE 754 also specifies decimal floating point arithmetic, which has never seen wide adoption. The reason why Octave and other programs use IEEE 754 binary floats is that they are ''fast'', because they are implemented in hardware or system libraries. Unless you are using very exotic hardware, Octave will use your computer's processor for basic floating point arithmetic.

Like death and taxes, rounding errors are a fact of life. You cannot avoid them. You can only move a rounding error from one part of a computation to another, or you can use more precision and delay the rounding error. One way to delay the rounding error is to use arbitrary precision arithmetic, which is inevitably slower as it has to be implemented in software instead of hardware. You may use the vpa function from the symbolic package for this purpose.

Another approach to the problem is interval arithmetic with the [[Interval package]]. Then, the exact result would always be enclosed by two binary floats. Again, this is slower since only the most basic interval arithmetic operations can be performed in hardware.

To learn more about floating point arithmetic, consult [http://en.wikipedia.org/wiki/Floating_point_arithmetic its Wikipedia article] or the classical reference [http://floating-point-gui.de/ What Every Computer Scientist Should Know About Floating Point Arithmetic].

 

You have probably forgotten to load the package. Use {{Codeline|pkg load package-name}} to load it. Most packages are no longer loaded automatically to avoid surprises. See reasoning on related FAQ [[FAQ#How_do_I_install_all_Octave_packages.3F|how do I install all Octave packages]]. If you want a specific package to be loaded by default at startup, consider adding the {{Codeline|pkg load}} command on your {{path|[[.octaverc]]}} file.

Users with this kind of problem should try to install/update their Intel OpenGL drivers for Windows or consider installing Mesa drivers from http://qt-project.org/wiki/Cross-compiling-Mesa-for-Windows

See also https://www.opengl.org/wiki/FAQ#Why_is_my_GL_version_only_1.4_or_lower.3F

which is not quite the same thing. There are still a number of differences between Octave and Matlab, however in general differences between the two are considered as bugs. Octave might consider that the bug is in Matlab and do nothing about it, but generally functionality is almost identical. If you find an important functional difference between Octave behavior and Matlab, then you should send a description of this difference (with code illustrating the difference, if possible) to http://bugs.octave.org.

Octave has limited support for nested functions. That is

          function y = foo (x)

            y = bar(x)

            function y = bar (x)

              y = ...;

            end

          end

          function y = foo (x)

            y = bar(x)

          function y = bar (x)

            y = ...;

The main difference with Matlab is a matter of scope. While nested functions have access to the parent function's scope in Matlab, no such thing is available in Octave, due to how Octave essentially “un-nests” nested functions.

* Some limitations on the use of function handles. The major difference is related to nested function scoping rules (as above) and their use with function handles.

* Matlab classdef object oriented programming is not yet supported, though work is underway in a branch of the development tree.

A large number of the Matlab core functions (ie those that are in the core and not a toolbox) are implemented, and certainly all of the commonly used ones. There are a few functions that aren't implemented, usually to do with specific missing Octave functionality (GUI, DLL, Java, ActiveX, DDE, web, and serial functions). Some of the core functions have limitations that aren't in the Matlab version. For example the sprandn function can not force a particular condition number for the matrix like Matlab can. Another example is that testing and the runtests function work differently in Matlab and Octave.

Matlab includes a "Just-In-Time" compiler. This compiler allows the acceleration of for-loops in Matlab to almost native performance with certain restrictions. The JIT must know the return type of all functions called in the loops and so you can't include user functions in the loop of JIT optimized loops. Octave doesn't have a JIT and so to some might seem slower than Matlab. For this reason you must vectorize your code as much as possible. The MathWorks themselves have a good document discussing vectorization at http://www.mathworks.com/support/tech-notes/1100/1109.html.

On a related point, there is no Octave compiler, and so you can't convert your Octave code into a binary for additional speed or distribution. There have been several aborted attempts at creating an Octave compiler. Should the JIT compiler above ever be implemented, an Octave compiler should be more feasible.

Up to Octave 2.9.9 there was no support for graphic handles in Octave itself. In the 3.2.N versions of Octave and beyond, the support for graphics handles is converging towards full compatibility. The patch function is currently limited to 2-D patches, due to an underlying limitation in gnuplot, but the experimental OpenGL backend is starting to see an implementation of 3-D patches.

There are no Matlab compatible GUI functions yet.  This might be an issue if you intend to exchange Octave code with Matlab users. There are a number of bindings from Octave to {{Forge|tcl-octave|Tcl/Tk}}, [http://octaviz.sourceforge.net/index.php? VTK] and {{Forge|zenity}} for example, that can be used for a GUI, but these are not Matlab compatible. Work on a Matlab compatible GUI is in an alpha stage in the QtHandles project, which may form part of a future release of Octave.

Octave itself includes no Simulink support. Typically the simulink models lag research and are less flexible, so shouldn't really be used in a research environment. However, some Matlab users that try to use Octave complain about this lack.

Octave includes an API to the Matlab MEX interface. However, as MEX is an API to the internals of Matlab and the internals of Octave differ from Matlab, there is necessarily a manipulation of the data to convert from a MEX interface to the Octave equivalent. This is notable for all complex matrices, where Matlab stores complex arrays as real and imaginary parts, whereas Octave respects the C99/C++ standards of co-locating the real/imag parts in memory. Also due to the way Matlab allows access to the arrays passed through a pointer, the MEX interface might require copies of arrays (even non complex ones).

Block comments denoted by {{Codeline|#{}} and {{Codeline|#&#125;}}  markers (or {{Codeline|%{}} and {{Codeline|%&#125;}}) are supported by Octave with some limitations. The major limitation is that block comments are not supported within [] or {}.

There are some differences in the mat v5 file format accepted by Octave. Matlab recently introduced the "-V7.3" save option which is an HDF5 format which is particularly useful for 64-bit platforms where the standard Matlab format can not correctly save variables. Octave accepts HDF5 files, but is not yet compatible with the "-v7.3" versions produced by Matlab.

Thanks to Daniel Kraft's 2011 Google Summer of Code project, Octave has a profiler since version 3.6.0. However, at the moment it only produces text output and has its own makeshift interface for hierarchical profiling.

Octave is a community project and so the toolboxes that exist are donated by those interested in them through [[Octave Forge]]. These might be lacking in certain functionality relative to the Matlab toolboxes, and might not exactly duplicate the Matlab functionality or interface.

==Short-circuit {{Codeline|&}} and {{Codeline|&#124;}} operators==

The {{Codeline|&}} and {{Codeline|&#124;}} operators in Matlab short-circuit when included in a condition (e.g. an {{Codeline|if}} or {{Codeline|while}} statement) and not otherwise. In Octave only the {{Codeline|&&}} and {{Codeline|&#124;&#124;}} short circuit. Note that this means that

            if (a | b)

              ...

            end

            t = a | b;

            if t

              ...

            end

have different semantics in Matlab. This is really a Matlab bug, but there is too much code out there that relies on this behavior to change it. Prefer the {{Codeline|&#124;&#124;}} and {{Codeline|&&}} operators in {{Codeline|if}} statements if possible. If you need to use code written for Matlab that depends on this buggy behavior, you can enable it since Octave 3.4.0 with the following command:

            do_braindead_shortcircuit_evaluation(1)

Note that the difference with Matlab is also significant when either argument is a function with side effects or if the first argument is a scalar and the second argument is an empty matrix. For example, note the difference between

            t = 1 | [];          ## results in [], so...

            if (t) 1, end       ## in if ([]), this is false.

            if (1 | []) 1, end   ## short circuits so condition is true.

            t = [1, 1] | [1, 2, 3];          ## error

            if ([1, 1] | [1, 2, 3]) 1, end  ## OK

Also Matlab requires the operands of {{Codeline|&&}} and {{Codeline|&#124;&#124;}} to be scalar values but Octave does not (it just applies the rule that for an operand to be considered true, every element of the object must be nonzero or logically true).

            if ([0, 1]) == if (all ([0, 1]))  ==>  i.e., condition is false.

            if ([1, 1]) == if (all ([1, 1]))  ==>  i.e., condition is true.

            if ([])

            if (all ([]))

because, despite the name, the {{Codeline|all}} is really returning true if none of the elements of the matrix are zero, and since there are no elements, well, none of them are zero. This is an example of [http://en.wikipedia.org/wiki/Vacuous_truth vacuous truth]. But, somewhere along the line, someone decided that {{Codeline|if ([])}} should be false. Mathworks probably thought it just looks wrong to have {{Codeline|[]}} be true in this context even if you can use logical gymnastics to convince yourself that "all" the elements of an empty matrix are nonzero. Octave however duplicates this behavior for {{Codeline|if}} statements containing empty matrices.

Matlab's solvers as used by the operators mldivide (\) and mrdivide (/), use a different approach than Octave's in the case of singular, under-, or over-determined matrices. In the case of a singular matrix, Matlab returns the result given by the LU decomposition, even though the underlying solver has flagged the result as erroneous. Octave has made the choice of falling back to a minimum norm solution of matrices that have been flagged as singular which arguably is a better result for these cases.

          function x = mldivide (A, b)

            m = rows(A);

            [Q, R, E] = qr(A);

            x = [A \ b, E(:, 1:m) * (R(:, 1:m) \ (Q' * b))]

          end

A numerical question arises: how big can the null space component become, relative to the minimum-norm solution? Can it be nicely bounded, or can it be arbitrarily big? Consider this example:

          m = 10;

          n = 10000;

          A = ones(m, n) + 1e-6 * randn(m,n);

          b = ones(m, 1) + 1e-6 * randn(m,1);

          norm(A \ b)

while Octave's minimum-norm values are around 3e-2, Matlab's results are 50-times larger. For another issue, try this code:

          m = 5;

          n = 100;

          j = floor(m * rand(1, n)) + 1;

          b = ones(m, 1);

          A = zeros(m, n);

          A(sub2ind(size(A),j,1:n)) = 1;

          x = A \ b;

          [dummy,p] = sort(rand(1,n));

          y = A(:,p)\b;

          norm(x(p)-y)

It shows that unlike in Octave, mldivide in Matlab is not invariant with respect to column permutations. If there are multiple columns of the same norm, permuting columns of the matrix gets you different result than permuting the solution vector. This will surprise many users.

Since the mldivide (\) and mrdivide (/) operators are often part of a more complex expression, where there is no room to react to warnings or flags, it should prefer intelligence (robustness) to speed, and so the Octave developers are firmly of the opinion that Octave's approach for singular, under- and over-determined matrices is a better choice than Matlab's.

Note that using try/catch combined with {{Codeline|rethrow (lasterror ())}} can not guarantee that global variables will be correctly reset, as it won't catch user interrupts with Ctrl-C. For example

          global a

          a = 1;

          try

            _a = a;

            a = 2

            while true

            end

          catch

            fprintf ('caught interrupt\n');

            a = _a;

            rethrow (lasterror());

          end

          global a

          a = 1;

          unwind_protect

            _a = a;

            a = 2

            while true

            end

          unwind_protect_cleanup

            fprintf ('caught interrupt\n');

            a = _a;

          end

Typing Ctrl-C in the first case returns the user directly to the prompt, and the variable ''a'' is not reset to the saved value. In the second case the variable ''a'' is reset correctly. Therefore Matlab gives no safe way of temporarily changing global variables.

    octave:1> function s = hello_string (to_who)

    > ## Say hello

    > if nargin<1, to_who = "World"; end

    > s = ["Hello ",\

    >      to_who];

    > endfunction

    octave:2> hello_string ("Moon")

    ans = Hello Moon

The double quote, {{Codeline|"}}, may be used to delimit strings, in addition to the single quote {{Codeline|'}}. See the previous example. Also, double-quoted strings include backslash interpretation (like C++, C, and Perl) while single quoted are uninterpreted (like Matlab and Perl).

Lines can be continued with a backslash, {{Codeline|\}}, in addition to three points {{Codeline|...}}. See the previous example.

    octave:1> [3 1 4 1 5 9](3)

    ans = 4

    octave:2> cos([0 pi pi/4 7])(3)

    ans = 0.70711

    octave:1> if ! strcmp (program_name, "octave"),

    >  "It's an error"

    > else

    >  "It works!"

    > end

    ans = It works!

    unwind_protect

      body

    unwind_protect_cleanup

      cleanup

    end_unwind_protect

Matlab can be made to do something similar with their {{Codeline|onCleanup}} function that was introduced in 2008a. Octave also has {{Codeline|onCleanup}} since version 3.4.0.

Similar to the do-while loop in C and C++, Octave allows a do-until loop which does not exist in Matlab

  x = 0

  do

    x += 1;

  until (x == 10)

Borrowed from [http://stackoverflow.com/q/26948776/3565696 other languages], [http://www.gnu.org/software/octave/doc/v4.0.1/Broadcasting.html octave broadcasting] allows easy and readable vectorization.

  f = (1:0.1:2);

  # put angular frequencies on the first dimension to prepare broadcasting

  omega = 2 * pi * f(:);

  # time is already on the second dimension (row vector)

  t = 0:0.02:2;

  # the resulting s will be a 2-dimensional array

  s = sin(omega .* t);

  # which can be displayed as

  pcolor(t, f, s)

  xlabel("t (s)")

  ylabel("f (Hz)")

  function mult = a(val)

          mult = val.*2;

  endfunction

  %!test

  %! assert (a(3), 6);

   octave:1> test a

Example code block can be part of function file in a similar manner as test functions. For example to run demo for function multinom of package specfun, use:

===Powerfull assert===