MATLAB save/load to file

SYNOPSIS

#include "matlab.h"
MAT     *m_load(FILE *fp, char **name)
MAT     *m_save(FILE *fp, MAT *A, char *name)
VEC     *v_save(FILE *fp, VEC *x, char *name)
double   d_save(FILE *fp, double d, char *name)

#include "matlab.h"
ZMAT    *zm_load(FILE *fp, char **name)
ZMAT    *zm_save(FILE *fp, ZMAT *A, char *name)
ZVEC    *zv_save(FILE *fp, ZVEC *x, char *name)
complex  z_save (FILE *fp, complex z, char *name)

DESCRIPTION

These routines read and write MATLAB\TM\ load/save files. This enables results to be transported between MATLAB and Meschach. The routine m_load() loads in a matrix from file fp in MATLAB save format. The matrix read from the file is returned, and name is set to point to the saved MATLAB variable name of the matrix. Both the matrix returned and name have allocated memory as needed. An example of the use of the routine to load a matrix A and a vector x is

MAT *A, *Xmat;
VEC *x;
FILE *fp;
char *name1, *name2;
  ......
if ( (fp=fopen("fred.mat","r")) != NULL )
{
    A    = m_load(fp,&name1);
    Xmat = m_load(fp,&name2);
    if ( Xmat->n != 1 )
    {  printf("Incorrect size matrix read in\n"); 
       exit(0);  }
    x = v_get(Xmat->m);
    x = mv_move(Xmat,0,0,Xmat->m,1,x,0);
}

The m_save() routine saves the matrix A to the file/stream fp in MATLAB save format. The MATLAB variable name is name. The v_save() routine saves the vector x to the file/stream fp as an x->dim${}\times1$ matrix (i.e.\ as a column vector) in MATLAB save format. The MATLAB variable name is name. The d_save() routine saves the double precision number d to the file/stream fp in MATLAB save format. The MATLAB variable name is name. The MATLAB save format can depend in subtle ways on the type of machine used, so you may need to set the machine type in machine.h. This should usually just mean adding a line to machine.h to be one of

#define MACH_ID INTEL           /* 80x87 format */
#define MACH_ID MOTOROLA        /* 6888x format */
#define MACH_ID VAX_D           /* VAX D format */
#define MACH_ID VAX_G           /* VAX G format */

to be the appropriate machine. The machine dependence involves both whether IEEE or non IEEE format floating point numbers are used, but also whether or not the machine is a ``little-endian'' or a ``big-endian'' machine.

BUGS

The m_load() routine will only read in the real part of a complex matrix. The routines are machine-dependent as described above.

SOURCE FILE: matlab.c, zmatlab.c