matGetFull, matPutFull

Purpose

Easiest way to get and put full Matrices into MAT-files.

C Synopsis

#include "matfile.h"
int matGetFull(fp,name,m,n,pr,pi)
  MATFile *fp;
  char *name;
  int *m;
  int *n;
  double **pr;
  double **pi;
int matPutFull(fp,name,m,n,pr,pi)
  MATFile *fp;
  char *name;
  int m;
  int n;
  double *pr;
  double *pi;

Fortran Synopsis

integer*4 function matGetFull(fp,name,m,n,pr,pi)
integer*4 function matPutFull(fp,name,m,n,pr,pi)
integer*4 fp,m,n,pr,pi
character*(*) name

Arguments

fp: MAT-file handle
m: row dimension
n: column dimension
name: name of Matrix to get or put to MAT-file
pi: pointer to imaginary part
pr: pointer to real part

Most MATLAB applications work only with full (non-sparse) Matrices. These routines provide an easy way to write a full Matrix into a MAT-file and to copy a full Matrix out of a MAT-file. They offer an alternative to matGetMatrix and matPutMatrix, which do not require use of the Matrix structure.

matGetFull reads the named Matrix from the MAT-file pointed to by fp and places the row dimensions, column dimensions, real array pointer, and imaginary array pointer into the locations specified by m, n, pr, and pi, respectively.

matGetFull returns 0 if successful, and 1 if the named variable can't be found, the named variable is not a full Matrix, or there is a file read error.

matGetFull allocates memory for the real and imaginary arrays using mxCalloc; use mxFree to return the memory when you are done.

If the Matrix is pure real, the imaginary pointer is NULL.

matPutFull writes the Matrix with dimensions m-by-n, real data pr, and imaginary data pi onto the MAT-file fp with the specified name.

If the Matrix does not exist on the MAT-file, it is appended to the end. If a Matrix with the same name already exists in the file, the existing Matrix is replaced with the new Matrix by rewriting the file.

Examples

Read the Matrix A from one MAT-file and write it out to another:

/* mattest4.c */ 
#include "mat.h"
main()
{
  MATFile *fp1, *fp2;
  int m, n
  double *pr, *pi;
  fp1 = matOpen("foo.mat","r");  
  fp2 = matOpen("foo2.mat","w");  
  matGetFull(fp1,"A",&m,&n,&pr,&pi);
  matPutFull(fp2,"A",m,n,pr,pi);
  matClose(fp1);
  matClose(fp2);
}

The Fortran code for this example is in mattest4.f:

  program main
  integer matOpen,matClose,matPutFull,matGetFull
  integer mf1, mf2, stat
  integer m, n, pr, pi
  mf1 = matOpen('foo.mat','r')
  mf2 = matOpen('foo2.mat','w')
  stat = matGetFull(mf1,'A',m,n,pr,pi)
  stat = matPutFull(mf2,'A',m,n,pr,pi)
  stat = matClose(mf1)
  stat = matClose(mf2)
c
  stop
  end

Write a simple real Matrix into a MAT-file. Name the Matrix A and the MAT-file foo.mat:

/* mattest5.c */ 
#include "mat.h"
static double Areal[6] = {1,2,3,4,5,6};
main()
{
  MATFile *fp;
  fp = matOpen("foo.mat","w");
  matPutFull(fp,"A",3,2,Areal,NULL);
  matClose(fp);
}

To test, first run mattest5; then go to MATLAB and enter:

load foo
A
A = 
  1  4
  2  5
  3  6

The Fortran code for this example is in mattest5.f: