convert a matrix with some zero to non zero matrix

for example I have this matrix: [ 1 2 3 ; 1 0 0 ; 0 0 0; 4 5 6] I want this matrix: [ 1 2 3 ; 1 nan nan ; 4 5 6 ] (also delete zero rows completely.)

Answer 1

[Chad Greene]

A = [ 1 2 3 ; 1 0 0 ; 0 0 0; 4 5 6]; 

% Find rows of all zeros: 
zerorows = sum(A,2)==0; 

% Discard rows of all zeros: 
A(zerorows,:) = []; 

% Find all remaining zeros in A: 
leftoverzeros = A==0; 

% Replace remaining zeros in A with NaNs: 
A(leftoverzeros) = nan

Answer 2

[Star Strider]

One approach:

M = [ 1 2 3 ; 1 0 0 ; 0 0 0; 4 5 6];                        % Original Matrix
M(M == 0) = NaN;                                            % Set: ‘0’ —> ‘NaN’
Mn = ~isnan(M);                                             % Values That Are Not ‘NaN’ Set To Logical ‘false’
M(sum(Mn,2) == 0,:) = []                                    % If All Columns Of Any Row Are ‘NaN’, Set That Row To ‘Empty’

M =

       1     2     3
       1 NaN     NaN
       4     5     6

EDIT — Added comment documentation for each line to describe what each line does. The code was not changed.

Answer 3

[Image Analyst]

I don't care for one of the two other answers since it uses sum() and thus will not be general in the case where you have negative numbers in your array. For example Chad's will remove the third row the A = [ 1 2 3 ; 1 0 0 ; 2, -1, -1; 4 5 6] even though the third row is not all zeros. It will however work for the example you gave of all positive integers. However Star's will work in all cases because the sum is after the isnan() and it's summing a map of where nan's are rather than summing the original matrix.

I offer the more general, robust way of using either any() or all() (your choice as to which to use as they are equivalent):

A = [ 1 2 3 ; 1 0 0 ; 0, 0, 0; 4 5 6]; 
% Find rows where all are 0
zeroRows = ~any(A, 2)
% or alternatively you could do it this way
zeroRows = all(A == 0, 2)
% Delete rows of all zeros:
A(zeroRows,:) = []; 
% Make any remaining zeros NaNs
A(A==0) = nan

[Chad Greene]

Great catch, Image Analyst!

Answer 4

[Azzi Abdelmalek]

M = [ 1 2 3 ; 1 0 0 ; 0 0 0; -1 -1  2]; 
out=arrayfun(@(x) strrep(x,0,nan),M(any(~~M,2),:))