矩阵加法跳过NaN值

Question

我想在跳过 NaN 值的同时对矩阵进行元素明智的加法。 MATLAB和Octave有 nansum ，但它在矩阵中进行列式加法。

让：

a = NaN * zeros(3)
b = ones(3)

我想要：

c = nan+(a, b)

和

c = b

和

d = nan+(a,a)

和

d = a

Answer 1

如果你沿n + 1维度连接你的n-d数组，你仍然可以使用nansum。

对于2D

% commands de-nested for readability. You can do this with a single line, of course
tmp = cat(3,a,b);
c = nansum(tmp,3);

一般情况

function out = nansumByElement(A,B)
%NANSUMBYELEMENT performs an element-wise nansum on the n-D arrays A and B
% A and B have to have the same size

% test input
if nargin < 2 || isempty(A) || isempty(B) || ndims(A)~=ndims(B) || ~all(size(A)==size(B))
error('please pass two non-empty arrays of the same size to nansumByElement')
end

% calculate output

nd = ndims(A); % get number of dimensions
% catenate and sum along n+1st dimension
out = nansum(cat(nd+1,A,B),nd+1);

Answer 2

a_fixed = a;
a_fixed(isnan(a)) = 0;
b_fixed = b;
b_fixed(isnan(b)) = 0;
c = a_fixed.+b_fixed;