Mercurial > hg > octave-lyh
view scripts/general/blkdiag.m @ 14872:c2dbdeaa25df
maint: use rows() and columns() to clarify m-files.
* gradient.m, interp1q.m, rat.m, tsearchn.m, image.m, imwrite.m, area.m,
contourc.m, hist.m, isocolors.m, isonormals.m, meshz.m, print.m, __bar__.m,
__go_draw_axes__.m, __interp_cube__.m, __marching_cube__.m, __patch__.m,
__print_parse_opts__.m, __quiver__.m, rose.m, shrinkfaces.m, stairs.m,
surfnorm.m, tetramesh.m, text.m, deconv.m, spline.m, intersect.m, setdiff.m,
setxor.m, union.m, periodogram.m, pcg.m, perms.m: Replace size (x,1) with
rows (x) and size(x,2) with columns(x).
author | Rik <octave@nomad.inbox5.com> |
---|---|
date | Tue, 17 Jul 2012 13:34:19 -0700 |
parents | f3d52523cde1 |
children |
line wrap: on
line source
## Copyright (C) 2000-2012 Daniel Calvelo ## ## This file is part of Octave. ## ## Octave is free software; you can redistribute it and/or modify it ## under the terms of the GNU General Public License as published by ## the Free Software Foundation; either version 3 of the License, or (at ## your option) any later version. ## ## Octave is distributed in the hope that it will be useful, but ## WITHOUT ANY WARRANTY; without even the implied warranty of ## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ## General Public License for more details. ## ## You should have received a copy of the GNU General Public License ## along with Octave; see the file COPYING. If not, see ## <http://www.gnu.org/licenses/>. ## -*- texinfo -*- ## @deftypefn {Function File} {} blkdiag (@var{A}, @var{B}, @var{C}, @dots{}) ## Build a block diagonal matrix from @var{A}, @var{B}, @var{C}, @dots{} ## All the arguments must be numeric and are two-dimensional matrices or ## scalars. If any argument is of type sparse, the output will also be ## sparse. ## @seealso{diag, horzcat, vertcat, sparse} ## @end deftypefn ## Author: Daniel Calvelo ## Modified by: William Poetra Yoga Hadisoeseno function retval = blkdiag (varargin) if (nargin < 1) print_usage (); endif if (! all (cellfun ("isnumeric", varargin))) error ("blkdiag: all arguments must be numeric"); endif ## Note: trailing singletons are automatically (correctly) ignored. if (! all (cellfun ("ndims", varargin) == 2)) error ("blkdiag: all arguments must be two-dimensional matrices"); endif ## size is an option for cellfun, but it's a bit different from ## calling size directly. tmp = cell2mat (cellfun (@size, varargin', "uniformoutput", false)); csz = cumsum ([0 0; tmp], 1); if (any (cellfun ("issparse", varargin))) retval = sparse (csz(end,1), csz(end,2)); else retval = zeros (csz(end,:)); endif for p = 1:nargin vp = varargin{p}; if (! isempty (vp)) retval((csz(p,1)+1):csz(p+1,1),(csz(p,2)+1):csz(p+1,2)) = vp; endif endfor endfunction ## regular tests %!assert (blkdiag (1,ones (2),1), [1,0,0,0;0,1,1,0;0,1,1,0;0,0,0,1]) %!assert (blkdiag ([1,2],[3,4],[5,6]), [1,2,0,0,0,0;0,0,3,4,0,0;0,0,0,0,5,6]) %!assert (blkdiag ([1,2],[3;4],[5,6]), [1,2,0,0,0;0,0,3,0,0;0,0,4,0,0;0,0,0,5,6]) %!assert (blkdiag ([1,2;3,4],[5,6,7]), [1,2,0,0,0;3,4,0,0,0;0,0,5,6,7]) ## tests involving empty matrices %!assert (blkdiag ([],[],[]), []) %!assert (blkdiag ([],[1,2;3,4],[],5,[]), [1,2,0;3,4,0;0,0,5]) %!assert (blkdiag (zeros (1,0,1),[1,2,3],1,0,5,zeros (0,1,1)), [0,0,0,0,0,0,0;1,2,3,0,0,0,0;0,0,0,1,0,0,0;0,0,0,0,0,0,0;0,0,0,0,0,5,0]); ## tests involving sparse matrices %!assert (blkdiag (sparse ([1,2;3,4]),[5,6;7,8]), sparse ([1,2,0,0;3,4,0,0;0,0,5,6;0,0,7,8])) %!assert (blkdiag (sparse ([1,2;3,4]),[5,6]), sparse ([1,2,0,0;3,4,0,0;0,0,5,6])) # sanity checks %!test %! A = rand (round (rand (1, 2) * 10)); %! assert (blkdiag (A), A);