fprintf for character above 127 uses 3 bytes

Hello Everyone, could somebody give me a hint why apply fprintf to a character beyond 127 would result a consumption of 3 bytes? Is this related to Matlab or the system? Thank you in advance.

Here is my test code, and further below is part of the results I got out of this code:

fid=fopen('testprintbeyond128','w')
for i=1:255
     fprintf(fid,'%s\t',num2str(i));       
     fprintf(fid,'%s\t',dec2hex(i)); 
     cc=fprintf(fid,'%s\t',char(i));
     fprintf(fid,'%d\n',cc)
end
        fclose(fid)

results:

123  7B  {  2
124  7C  |  2
125  7D  }  2
126  7E  ~  2
127  7F    2
128  80    3
129  81    3
130  82    3

Answer 1

[Guillaume]

Once you've opened your text file, check its encoding with

[~, ~, ~, encoding] = fopen(fid)

You will find that it is UTF-8. In UTF-8 character codes U+0000 between U+007F are encoded on one byte and character codes between U+0080 and U+07FF are encoded on two bytes. Add one byte for the tab (U+0009) and what you're seeing is normal.

[MarsMat]

Hi Guillaume, Thank you for your comments. Then I'd like to ask if there is a way to write characters beyond 127 with 1 byte? I got a file which was saved in such a way, But I can not write back in the same way. So to speak, if I use "fread" to read in the file byte by byte, at some of them it is a number larger than 127. As shown in the code, I expect it to be 1 byte character by 128, but if I read it back, it became a 2 bytes (fread get back 2 numbers). How would I solve this?

fid=fopen('test','w')
fprintf(char(128))
fclose(fid)

fid=fopen('test','r')
test=fread(fid)
fclose(fid)

[Walter Roberson]

See http://www.mathworks.com/matlabcentral/answers/275091-fprintf-and-fread-character-beyond-ascii-in-1-byte

[Guillaume]

@MarsMat, two things:

If you use fprintf then really you should read the data back with fscanf. This would handle all these encoding issues transparently. While, we're at it, I would open the file with 'wt' and 'rt' to make sure end of lines are handled properly.

To have character from 128 to 255 stored as one byte, you need to explicitly specify an adequate character encoding when you fopen the file. Only you can tell which one is the most appropriate, 'ISO-8859-1' is the closest to unicode but depending on where you're located others may be more appropriate.