How to arrange a matrix in descending order w.r.t rows?

If we have a large matrix and we want to arrange it in descending order i.e., the largest row shoud come on top, then 2nd largest row comes as 2nd row, then 3rd largest as 3rd row and so on but the columns are not disturbed. I do it like this:

clear all; clc
a=[1 2 3 4;1.1 2.1 3.1 4.1;1.2 0 3.2 4.2];
b=sort(a,"descend");
[a b]

But in this the columns are also disturbed.

Answer 1

[the cyclist]

Probably use the sortrows function. I'd be more specific, but it is unclear to me what you mean by "largest row". For example, which row is larger?

M = [2 7 11 19;
     3 5 13 17];

[Sadiq Akbar]

Thanks a lot dear the cyclist for your kind response. Actually I have data like this:

fval=[2.11 2.10 2.13 2.18 2.09];

a=[1    2    3    4;
   1.1  2.1  3.1  4.1;
   1.2  0    3.2  4.2;
   1.01 2.01 3.01 4.01;
   1.1  2.9  3.1  4.2];

Here the 1st value in 'fval' corresponds to 1st row of 'a'. The 2nd value in 'fval' corresponds to the 2nd row of 'a' and so on. 1st I want to arrange 'fval' in descending ordr. I do it like this:

fval1=sort(fval,'descend');

Then I want to arrange the matrix 'a' in the same order as is the new fval1. so I do it like this:

[fval1 ind]=sort(fval,'descend');
b=a(ind,:);

But I am not sure whether I am doing it correctly or not? Because when I find the norm of 'b', its values are not in descending order.

[the cyclist]

I'm not sure I understand. I ran your code here (and wrote a in a way that makes the rows clearer).

Each row stays intact, but is moved up or down, based on the sorting you did. This isn't what you want? Maybe you could write out the output you expected to get?

fval=[2.11 2.10 2.13 2.18 2.09];

a=[1    2    3    4;
   1.1  2.1  3.1  4.1;
   1.2  0    3.2  4.2;
   1.01 2.01 3.01 4.01;
   1.1  2.9  3.1  4.2];

[fval1 ind]=sort(fval,'descend');
b=a(ind,:)

b = 5×4

1.0100 2.0100 3.0100 4.0100 1.2000 0 3.2000 4.2000 1.0000 2.0000 3.0000 4.0000 1.1000 2.1000 3.1000 4.1000 1.1000 2.9000 3.1000 4.2000

[Sadiq Akbar]

Thanks a lot for your kind response. May be I am unable to communicate my question properly. Look I run this code:

clear all; clc
fval=[2.11 2.10 2.13 2.18 2.09];
a=[1    2    3    4;
   1.1  2.1  3.1  4.1;
   1.2  0    3.2  4.2;
   1.01 2.01 3.01 4.01;
   1.1  2.9  3.1  4.2];
[fval1 ind]=sort(fval,'descend');
b=a(ind,:);

for ii=1:length(a)
Na(ii)=norm(a(ii,:));
Nb(ii)=norm(b(ii,:));
end
Both=[Na' Nb']

Now when we look at the 2nd column of Both, the values in that column are not in the descending order. I want to make them in descending order and then arrange the matrix 'a' accordingly and then find the row which corresponds to the minimum value of Nb. But I am confused.

Answer 2

[the cyclist]

Answering here, based on your comment to my other answer.

It seems that you just want to sort the matrix a according to its norm. It also seems that the vector fval has absolutely nothing to do with the norm. So why are you sorting by fval?

This code will create the matrix b, which it seems is what you want.

% Original matrix
a = [1    2    3    4;
     1.1  2.1  3.1  4.1;
     1.2  0    3.2  4.2;
     1.01 2.01 3.01 4.01;
     1.1  2.9  3.1  4.2];

% Calculate the norm of each row of a
for ii = 1:height(a)
    Na(ii) = norm(a(ii,:));
end

% Find the sorting order index
[~, ind]=sort(Na,'descend');

% Define the matrix b, which is a sorted by the norm
b = a(ind,:);

% Calculate the norm of each row of b
for ii = 1:height(a)
    Nb(ii) = norm(b(ii,:));
end

% Show that b is sorted by the norm
Both = [Na' Nb']

Both = 5×2

5.4772 6.0721 5.6604 5.6604 5.4148 5.4955 5.4955 5.4772 6.0721 5.4148

[Sadiq Akbar]

I tried to modify my code according to yours but I am stuck. I am going to share my data here along with the code. The data is in the attachment and the code is below:

clear all; clc
load 2sn0dB
fval=one;
Error=u-two;% two is my matrix b

for ii=1:length(Error)
NError(ii)=norm(Error(ii,:));%Norm of the error
end

% Arranging the column vector 'NError' in descending order
[NError,ind]=sort(NError','descend');

% Arranging matrix two according to indices of NError
two2=two(ind,:);

In this two is my matrix b of size 100 x 4. I want to find the error between each of its row with vector u. Then I want to find the norm of each row of matrix Error. After that I want to arrange those norms into descending order. Then I want to arrange the matrix two according to the indices of those sorted norms. Then I want to decrease the error between u and matrix two but how? And then I want to update the matrix two such that their values do not deviate too much from u. But I don't know whether I am doing it correctly or wrongly?