compare two values and chosing the higher one

user = 2;
subcarrier = 4;
randn(user,subcarrier)

If i run this i am getting

   ans =

     -0.1623   -0.5320   -0.8757   -0.7120
     -0.1461    1.6821   -0.4838   -1.1742

from the values i want ans to be displayed as

   -0.1623   0      -0.8757     -0.7120
   0         1.6821  0           0.

how can i modify the code to get this output

[Prabha Kumaresan]

0          0         0       -0.7120
-0.1461    1.6821    -0.4838  0

for this output i want to modify the code

Answer 1

[KSSV]

A = [-0.1623   -0.5320   -0.8757   -0.7120
   -0.1461    1.6821   -0.4838   -1.1742];

B = [-0.1623   0      -0.8757     -0.7120
 0         1.6821  0           0.] ;

C = zeros(size(A)) ;

[val,idx] = max(abs(A)) ;
% [I_row, I_col] = ind2sub(size(A),idx) ;
% C(I_row,I_col) = val ;
for i = 1:size(A,2) 
    [val,idx] = max(abs(A(:,i))) ;
    C(idx,i) = val ;
end

[Prabha Kumaresan]

Could you tell me how to modify the code if A has 10 rows and 64 columns of values instead of [-0.1623 -0.5320 -0.8757 -0.7120 -0.1461 1.6821 -0.4838 -1.1742];

Answer 2

[Walter Roberson]

user = 2;
subcarrier = 4;
temp = randn(user,subcarrier);

temp( bsxfun(@lt, temp, max(temp)) ) = 0

Note: your requirement is not defined in the case where there are duplicate maximum values.

[Walter Roberson]

My code is already set up to give only one non-zero entry per column. (All of the entries in a column could be 0 if it happened that randn produced an exact 0!)

Answer 3

[Andrei Bobrov]

a = [     -0.1623   -0.5320   -0.8757   -0.7120
-0.1461    1.6821   -0.4838   -1.1742];

out = (a == max(a)).*a;

or

out = a.*bsxfun(@eq,max(a),a);