MEX question: How to set scalar and array values within plhs[0] struct?

I've been pulling my hair out about this one for a while. I have a C struct shown below, and am trying to convert it to be the exact same type of setup to be returned within plhs[0]. However when I print the contents of the first value of the plhs struct, I get -nan. I'd really apreciate the help to figure out why this isn't working correctly.

Here's the C struct:

typedef struct {
  boolean_T nL;
  int lID;
  int mID;
  int mT;
  int tID;
  int tT;
  double t0;
  double PIP[3];
  double UFD[3];
} struct1_T;

Here's my C code to set the "nL" boolean within plhs[0]:

    int dims[2];
    dims[0] = 1;
    dims[1] = 1;
    mwIndex m = 0;
    const char *names[] = {"nL","lID","mID","mT","tID","tT","t0","PIP","UFD"};
    plhs[0] = mxCreateStructArray(2, dims, 9, names);
    int conv = data[0].nL; //data is a struct1_T struct shown above
    double temp = conv;
    mexPrintf("val: %f\n", temp);
    mxSetField(plhs[0], 0, names[0], mxCreateDoubleScalar(temp));
    double *print = mxGetField(plhs[0],m,"nL");
    mexPrintf("Output_0: %f\n" print[0]);

And here is the output I get when I run it:

val: 0.000000
Output_0: -nan

Answer 1

[James Tursa]

mxGetField returns a pointer to mxArray, not pointer to double. You need to get at the data with the appropriate function. E.g.,

double *print = mxGetPr(mxGetField(plhs[0],m,"nL")); // R2017b memory model

or

double *print = mxGetDoubles(mxGetField(plhs[0],m,"nL")); // R2018a memory model

The above is quick and dirty code. Normally one would check the result of mxGetField first to make sure it is not NULL and that it contains a non-empty double variable before calling mxGetPr or mxGetDoubles.

Didn't the compiler give you a warning about your current code converting a (mxArray *) type to a (double *) type? In C++ that line would have generated a compile error.

[Jerry Painter]

D'oh. Why didn't I catch that? Now I'm having a different problem... plhs[0] for t0 isn't returning the correct value.

    mexPrintf("Struct_t0: %d\n",data[0].t0);
    mxSetField(plhs[0],6,"t0",mxCreateDoubleScalar(data[0].t0));
    mexPrintf("plhs[0]_t0: %f\n",mxGetPr(mxGetField(plhs[0],6,"t0")));

Output:

Struct_t0: 22
plhs[0]_t0: 0.000000

[James Tursa]

You are using a struct index of 6 now instead of 0. What are the values of dims[0] and dims[1]? Are you writing/reading beyond the end of the array?

Also, you are printing the result of the mxGetPr() call using a floating point format. That's not going to work. You need to dereference that pointer. E.g.,

    mexPrintf("plhs[0]_t0: %f\n",*mxGetPr(mxGetField(plhs[0],6,"t0")));