[Previous] [Next]

XPRSwriteprob

WRITEPROB

Purpose

Writes the current problem to an MPS or LP file.

Synopsis

int XPRS_CC XPRSwriteprob(XPRSprob prob, const char *filename, const char *flags);

WRITEPROB [-flags] [filename]

Arguments

prob  

The current problem.

filename  

A string of up to 200 characters to contain the file name to which the problem is to be written. If omitted, the default problem_name is used with a .mat extension, unless the l flag is used in which case the extension is .lp.

flags  

Flags, which can be one or more of the following:

p	full precision of numerical values;
o	one element per line;
n	scaled;
s	scrambled vector names;
l	output in LP format;
x	output MPS file in hexadecimal format.

Example 1 (Library)

The following example outputs the current problem in full precision, LP format with scrambled vector names to the file problem_name.lp.

XPRSwriteprob(prob, "", "lps");

Example 2 (Console)

WRITEPROB -p C:myprob

This instructs the Optimizer to write an MPS matrix to the file myprob.mat on the C: drive in full precision.

Further information

If XPRSloadlp, XPRSloadglobal, XPRSloadqglobal or XPRSloadqp is used to obtain a matrix then there is no association between the objective function and the N rows in the matrix and so a separate N row (called __OBJ___) is created when you do an XPRSwriteprob (WRITEPROB). Also if you do an XPRSreadprob (READPROB) and then change either the objective row or the N row in the matrix corresponding to the objective row, you lose the association between the two and the __OBJ___ row is created when you do an XPRSwriteprob (WRITEPROB). To remove the objective row from the matrix when doing an XPRSreadprob (READPROB), set KEEPNROWS to -1 before XPRSreadprob (READPROB).

The hexadecimal format is useful for saving the exact internal precision of the matrix.

Warning: If XPRSreadprob (READPROB) is used to input a problem, then the input file will be overwritten by XPRSwriteprob (WRITEPROB) if a new filename is not specified.