1function [vnz,rnz,parent,c,leftmost,p,q] = cs_sqr (A) %#ok 2%CS_SQR symbolic sparse QR factorization. 3% [vnz,rnz,parent,c,leftmost,p] = cs_sqr(A): symbolic QR of A(p,:). 4% [vnz,rnz,parent,c,leftmost,p,q] = cs_sqr(A) computes the symbolic QR 5% factorization of A(p,q). The fill-reducing ordering q is found via 6% q = cs_amd(A,3). 7% 8% vnz is the number of entries in the matrix of Householder vectors, V. 9% rnz is the number of entries in R. parent is elimination tree. 10% c(i) is the number of entries in R(i,:). leftmost(i) = min(find(A(i,q))). 11% p is the row permutation used to ensure R has a symbolically zero-free 12% diagonal (it can be larger than m if A is structurally rank deficient). 13% q is the fill-reducing ordering, if requested. 14% 15% Example: 16% Prob = ssget ('HB/ibm32') ; A = Prob.A ; 17% [vnz, rnz, parent, c, leftmost, p, q] = cs_sqr (A) ; 18% cspy (A (p,q)) ; 19% 20% See also CS_AMD, CS_QR. 21 22% Copyright 2006-2012, Timothy A. Davis, http://www.suitesparse.com 23 24error ('cs_sqr mexFunction not found') ; 25