MATLAB: Passing unicode string from C mex function to Matlab

MATLABmexmxcreatestringunicodeutf16utf8

Hello, I need to send a unicode string from my C mex function to Matlab. At the moment, I try to use mxCreateString to create the variable to return, but unfortunately, I do not get the correct string in Matlab. The string in my C mex function is UTF8 encoded. As far as I can see, the mxCreateString takes a char* only. Does this mean, that mxCreateString can not handle unicode strings? Is there any other way to return unicode strings to Matlab?

Am I correct that Matlab itself uses UTF16 to store strings?

Thank you in advance,

Christian

Best Answer

Hi,

at least the doc states that the mxChar type is 2 byte:

http://www.mathworks.com/help/matlab/apiref/mxchar.html

What happens when you use mxCreateCharArray together with a mxGetData and a memcopy? Or you use a loop to write the data in and cast accordingly?

mxArry *tmp = mxCreateCharArray(...)
mxChar *charData = (mxChar *)mxGetData(tmp);
        for(j=0; j < nelem_tmp; j++) {
                  charData[j] = (mxChar)Your_string[j]
}

Related Solutions

MATLAB: UTF-8 strings in MEX-files

A solution when using C++11:

I found this answer on StackOverflow: http://stackoverflow.com/a/38383389 It turns out C++11 has built-in functions for converting between various Unicode encodings.

#include "mex.h"
#include <algorithm>
#include <string>
#include <locale>
#include <codecvt>
...
std::string u8str = u8"µm²";
std::u16string u16str = std::wstring_convert< std::codecvt_utf8_utf16< char16_t >, char16_t >{}.from_bytes( u8str );
mwSize sz[ 2 ] = { 1, u16str.size() + 1 }; // +1 to include terminating null character
mxArray* mxstr = mxCreateCharArray( 2, sz );
std::copy( u16str.begin(), u16str.end() + 1, mxGetChars( mxstr )); // again +1 for terminating null character

MATLAB: Passing a String from Matlab to a Fortran Mex Function

Hi Rick - the revord.F example does a decent job of describing the steps to read the input string. It can be opened in your editor as

 edit([matlabroot '/extern/examples/refbook/revord.F']);

The code that follows is a variation of the this function. Note that I had problems with the call mexErrMsgIdAndTxt ... calls crashing MATLAB, so I just replaced them with a mexPrintf equivalent.

 #include "fintrf.h"
 C     Gateway routine
       subroutine mexFunction(nlhs, plhs, nrhs, prhs)
 C     mexFunction arguments:
       mwPointer plhs(*), prhs(*)
       integer nlhs, nrhs
 C     Function declarations:
       mwPointer mxCreateString
       mwPointer mxGetString 
       mwPointer mxGetM, mxGetN
       integer mxIsChar
 C     Array information:
       mwPointer mrows, ncols
 C     Arguments for computational routine:
       mwSize  maxbuf
       parameter(maxbuf = 100)
       character*120 line
       integer*4 k
       character*100 input_buf 
       mwPointer  strlen
 C     Check for proper number of arguments. 
       if(nrhs .ne. 1) then
          write(line,*) 'error: incorrect number of input arguments'
          k=mexPrintf(line//achar(13)) 
          return
 C     The input must be a string.
       elseif(mxIsChar(prhs(1)) .ne. 1) then
          write(line,*) 'error: input argument is not a string'
          k=mexPrintf(line//achar(13)) 
          return
       endif
 C     Get the size of the input string.
       mrows = mxGetM(prhs(1))
       ncols = mxGetN(prhs(1))
 C     Get the length of the input string and validate.
       strlen = mrows*ncols
       if (strlen .gt. maxbuf) then
          write(line,*) 'error: input is greater than max str size'
          k=mexPrintf(line//achar(13)) 
          return
       endif
 C     Get the string.
       status = mxGetString(prhs(1), input_buf, maxbuf)
       write(line,*) 'str = ', input_buf
       k=mexPrintf(line//achar(13)) 
       return
       end

The code checks to make sure that there is only one string input. Note how the input string can only have a fixed number of characters. I saved the above to a file called readString.F and compiled as mex readString.F. In the Command Window, I called this function as

 >> readString(1)
 error: input argument is not a string    
 >> readString('hello',2)
 error: incorrect number of input arguments   
 >> readString('hello')
 str = hello      
 >> readString('This is a place of business')
 str = This is a place of business

See the attached for a working copy of this code.

Best Answer

Related Solutions

MATLAB: UTF-8 strings in MEX-files

MATLAB: Passing a String from Matlab to a Fortran Mex Function

Related Question