Can fprintf UTF-8 text, but cannot use xmlwrite

23 views (last 30 days)
I have a structure as:
q.txt(1:100) % contains 100 UTF-8 arabic words
I can dump a range to a file as:
fprintf(fid,'%s', char(strjoin(q.txt(30:40))));
The resultant file is UTF-8 encoded " with BOM " and can be viewed by any text editor such as NotePad++. But when I try to store my text in xml, I get UTF-8 encoding " without BOM ". Notepade++ also detects its encoding as: ANSI as UTF-8. No UTF-8 arabic characters are written.
docNode = com.mathworks.xml.XMLUtils.createDocument('qdoc');
toc = docNode.getDocumentElement;
toc.setAttribute('version','0.01');
simitem = docNode.createElement('simitem');
simitem.appendChild(docNode.createTextNode(char(strjoin(q.txt(30:40)))));
toc.appendChild(simitem);
xmlwrite('info.xml',docNode);
The produced xml file does not properly show arabic strings, the encoding is not correct. But other tags and english attributes show normally.

Answers (1)

Noam Greenboim
Noam Greenboim on 21 May 2020
Many Matlab functions don't add BOM to the UTF-8 files, although the encoding was set to UTF-8 (I think it's the default). This includes writetable, writecell and more.
Therefore, you need to add it manually.
I had the same issue, and I made a work-around -
Usage:
BOM('info.xml', 'UTF-8')
This should add the BOM to your file. Note that it returns the currect BOM (before the change), if it exists.
The idea behind the hack is to edit the file as a binary, and then add the 3 characters BOM. They are invisible if you just use a text editor.
  2 Comments
Walter Roberson
Walter Roberson on 21 May 2020
The UTF-8 recommendation is not to add BOM.
Noam Greenboim
Noam Greenboim on 22 May 2020
Maybe. But when you open it with Excel, for example, it would mix-up the non-English text.

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!