How do I get my MATLAB editor to read UTF-8 characters? UTF-8 characters in blank squares in editors, but in the command window and workspace works fine.

I have a project, which is commented in UTF-8 characters.

I tried changing the system locale on my Windows 10, however MATLAB editor is not recognizing UTF-8 characters(in blank squares). I'm not sure what to do here.

If I open the same .m file in text editor, it works fine.

How do I get my MATLAB editor to read UTF-8? Thank you

feature('DefaultCharacterSet')
feature('locale')

ans =

UTF-8

ans = 

                 ctype: 'en_US.windows-1252'
               collate: 'en_US.windows-1252'
                  time: 'en_US.windows-1252'
               numeric: 'en_US_POSIX.windows-1252'
              monetary: 'en_US.windows-1252'
              messages: 'en_US.windows-1252'
              encoding: 'windows-1252'
      terminalEncoding: 'windows-949'
           jvmEncoding: 'Cp1252'
                status: 'MathWorks locale management system initialized.'
               warning: 'System locale setting, ko_KR, is different from user locale setti…'

[Tom Anderson]

If you find out the answer, let me know! But plenty of other comments around make it seem that UTF-8 just doesn't work in the MATLAB built-in editor in Windows 10.

Personally, I've been using Notepad++. By 2014b, UTF-8 works fine except for in the IDE/editor. I can do almost everything with the .m file, e.g. run and debug in MATLAB, make graphs. You can even view the file in the editor as long as you don't mind seeing a bit of messy encoding errors. But don't ever save the file with the built-in editor! It screws up the encoding. From time to time a UTF-8 file of mine gets screwed up by the editor, so always have a backup copy!

As far as I can tell, the problem is that the MATLAB editor saves files with ANSI encoding, not UTF-8. Maybe one day they will allow us to control this in the preferences for the Editor/Debugger.

[Hongyu Shi]

I'm having the same issue for a long time. It seems that sometimes it can handle Chinese characters when I change my system into Chinese default language. But in English system they just becomes a lot of question marks.

[Keith Hooks]

I'm having the same problem. It seems that editing the lcdata.xml file to change the encoding of en_US has no effect at all. It stays windows-1252 no matter what I change it to in the file.

Answer 1

[Jinghao Lei]

I have a very tricky way to solve this problem. And it seems works. In my case, (windows matlab 2016b x64)

feature('locale')

always output below even I have modified lcdata.xml

ctype: 'zh_CN.GBK'
...

so, I delete this in lcdata.xml (in codeset)

<encoding name="GBK">
    <encoding_alias name="936">
</encoding>

then I change following

<encoding name="UTF-8">
    <encoding_alias name="utf8"/>
</encoding>

to

<encoding name="UTF-8">
    <encoding_alias name="utf8"/>
    <encoding_alias name="GBK"/>
</encoding>

The point is cheat matlab GBK is just alias of utf8

[Oleg Komarov]

It's hacky but it works! Until TMW introduces a reliable way to set the encoding, this is the way to go.

[Francis Gagnon]

Indeed, it works ! Finally a solution to use another encoding in the MATLAB Editor.

Please, The MathWorks, include a way to switch char encoding on the fly in the MATLAB Editor! This feature should be there since a long time ago...

[Yueyan Li]

For MATLAB2017a users, as noticed there is empty in lcdata.xml except some comments. So we should first rename the lcdata_utf8.xml to lcdata.xml, then we do what Jinghao Lei said.

[Martin Herrerias Azcue]

I tried to do the opposite on Linux (I work on a collaborative project under both platforms) without success; but batch-converting everything to UTF-8 and cheating MATLAB on Windows to work with UTF-8 works like a charm...

I still have to add:

feature('DefaultCharacterSet','UTF-8');

to my main script for things to display properly on GUI elements when running under Windows.

[rubinius]

The editor shows correct characters in UTF-8. However, When I type

help myfunction

or view myfunction in the help browser, the function comments coded in UTF-8 does not show properly.

Is there a way to resolve this problem?

[Walter Roberson]

Peng Chen comments to rubinius:

I want to know the way to resolve the problem, too.

[Walter Roberson]

Peng Chen:

Which MATLAB release are you using, and which operating system?

[Isidora Timkov-Glumac]

@Jinghao Lei I have the same version of MatLab as you, but my ctype is 'en_US_POSIX.US-ASCII'. How should I modify the lcdata file so that this hack works for me? Thank you!

[Simon Diehl]

This solution worked for me as welll. I'm using Matlab R2023b on Windows Server 2019. My encoding was windows-1252.

This is not necessary on newer operating systems. I don't experience this problem with R2023b on Windows 10.

My steps where:

1. Go to Program Files\MATLAB\R2023b\bin
2. Rename lcdata.xml to lcdata_old.xml
3. Copy lcdata_utf8.xml and rename it to lcdata.xml
4. Open the file and go to section codeset <!-- Codeset entry -->
5. Comment out <encoding name="windows-1252" ...
6. Go to <encoding name="UTF-8"> and add the alias <encoding_alias name="1252">

The final sections look like this:

<!-- <encoding name="windows-1252" jvm_encoding="Cp1252">
    <encoding_alias name="1252"/>
</encoding> -->

<encoding name="UTF-8">
    <encoding_alias name="utf8"/>
    <encoding_alias name="1252"/>
</encoding>

And the result is:

feature("locale")

ans = 

  struct with fields:

               ctype: 'de_DE.UTF-8'
             collate: 'de_DE.UTF-8'
                time: 'de_DE.UTF-8'
             numeric: 'en_US_POSIX.UTF-8'
            monetary: 'de_DE.UTF-8'
            messages: 'de_DE.UTF-8'
            encoding: 'UTF-8'
    terminalEncoding: 'IBM850'
         jvmEncoding: 'UTF-8'
              status: 'MathWorks locale management system initialized.'
             warning: ''
             

The editor now works as expected. Thank you very much for the solution @Jinghao Lei

Answer 2

[Michael Cappello]

% read in the file
fID = fopen(filename, 'r', 'n', 'UTF-8');
bytes = fread(fID);
fclose(fID);

The data read from the file can then be converted into Unicode characters, like so:

unic = native2unicode(bytes, 'UTF-8');

if you want, clear the Carriage Returns, set the Line Feeds to a space

unic(unic == 10) = []; unic(unic == 13) = ' ';
disp(unic'); % display the Unicode text