Fixed size matrix file I/O (original) (raw)

Next: Conversion from general to Up: Fixed size matrices Previous: Fixed size matrix decompositions Contents

Fixed size matrix file I/O

Gandalf supports both ASCII and binary format file I/O of matrices. Both formats use standard FILE * file streams. ASCII format is obviously more convenient to use, while binary format is more compact and guarantees no loss of precision when the data is read. To print a $3\times 4$ matrix in ASCII format, use

  Gan_Bool gan_mat34_fprint ( FILE *fp, Gan_Matrix34 *p,
                              const char *prefix, int indent, const char *fmt );

prefix is a prefix string to print before the matrix itself, indentis the number of spaces to indent the matrix by, and fmt is a format string to use when printing the matrix, e.g. "%f". So for example

  FILE *pfFile;

  pfFile = fopen ( "/tmp/matrices", "w" );
  gan_mat34_fill_q ( &m34A, 1.0,  2.0,  3.0,  4.0,
                            5.0,  6.0,  7.0,  8.0,
                            9.0, 10.0, 11.0, 12.0 );
  gan_mat34_fprint ( pfFile, &m34A, "Example matrix", 3, "%f" );

will print the output

     Example matrix
      1.000000 2.000000 3.000000 4.000000
      5.000000 6.000000 7.000000 8.000000
      9.000000 10.000000 11.000000 12.000000

to the file "/tmp/matrices". There is also a version gan_mat34_print() for printing to standard output:

  Gan_Bool gan_mat34_print ( Gan_Matrix34 *p,
                             const char *prefix, int indent, const char *fmt );

The corresponding input function is

  Gan_Bool gan_mat34_fscanf ( FILE *fp, Gan_Matrix34 *p,
                              char *prefix, int prefix_len )

which reads the matrix from the file stream fp into the $3\times 4$ matrix pointer p. It also reads the prefix string (up to the specified maximum length prefix_len), which can be compared with the expected prefix string to check for consistency.

Binary file I/O is handled by the functions gan_mat34_fwrite()and gan_mat34_fread(). To write a $3\times 4$ matrix in binary format use

  Gan_Bool gan_mat34_fwrite ( FILE *fp, Gan_Matrix34 *p, gan_ui32 magic_number );

The magic_number takes the same role as the prefix string in gan_mat34_fprint(), and is written into the file so that it can be used later to identify the matrix when it is read back using

  Gan_Bool gan_mat34_fread ( FILE *fp, Gan_Matrix34 *p, gan_ui32 *magic_number );

There are similar functions gan_mat33_fprint(),gan_mat33_print(), gan_mat33_fscanf() for ASCII I/O of $3\times 3$ matrices, and gan_mat33_fwrite(), gan_mat33_fread()for binary I/O of $3\times 3$ matrices. Functions for symmetric and triangular matrices follow the same pattern.

Error detection: The I/O routines return a boolean value, returningGAN_TRUE on success, GAN_FALSE on failure, invoking the Gandalf error handle in the latter case.

Next: Conversion from general to Up: Fixed size matrices Previous: Fixed size matrix decompositions Contents

2006-03-17