for loop does not iterate

I am trying to get used to matlab, but I am stuck.

I'm not able to iterate through my x_roots and separate real roots form complex ones

This is my code:

clear all
close all
syms n

w = @(x) 0.01.*x.^4-0.04.*x.^3+0.02.*x.^2+0.04.*x-0.03;
w_vec = [0.01 0.04 0.02 0.04 0.03];

x = -2:.2:4;
w_y = w(x);

x_roots = roots(w_vec)

x_roots = 
  -3.6921 + 0.0000i
   0.2003 + 1.0521i
   0.2003 - 1.0521i
  -0.7084 + 0.0000i

x_realroots = [];

for i=x_roots
    ~round(imag(i))
    if (~round(imag(i)))
        round(imag(i))
        x_realroots = [x_realroots i];
    else
        continue;
    end
end

ans = 4×1 logical array
   1
   0
   0
   1

x_realroots

x_realroots =

     []

So yeah, I just wanted to get real roots, but now I'm learning basic matlab.

Please explain to me why i cannot iterate through x_roots.

Answer 1

[Voss]

"Please explain to me why i cannot iterate through x_roots."

A for loop in MATLAB iterates over the columns of what you give it. Here are some simple examples to illustrate what I mean:

Example 1: row vector

x = 1:3      % a row vector with 3 elements (i.e., 3 columns)

x = 1×3

1 2 3

count = 0;
for ii = x   % for loop iterates 3 times
    count = count+1;
    disp("Iteration: " + count)
    disp("ii = ")
    disp(ii) % ii is a column of x each time (each column of x is a single element)
end

Iteration: 1

ii = 

     1

Iteration: 2

ii = 

     2

Iteration: 3

ii = 

     3

Example 2: matrix:

x = magic(3) % a matrix with 3 columns

x = 3×3

8 1 6 3 5 7 4 9 2

count = 0;
for ii = x   % for loop iterates 3 times
    count = count+1;
    disp("Iteration: " + count)
    disp("ii = ")
    disp(ii) % ii is a column of x each time
end

Iteration: 1

ii = 

     8
     3
     4

Iteration: 2

ii = 

     1
     5
     9

Iteration: 3

ii = 

     6
     7
     2

Example 3: column vector:

x = [1;2;3]  % a column vector with 3 elements

x = 3×1

1 2 3

count = 0;
for ii = x   % for loop iterates 1 time (x has only one column)
    count = count+1;
    disp("Iteration: " + count)
    disp("ii = ")
    disp(ii) % ii is the entirety of x
end

Iteration: 1

ii = 

     1
     2
     3

In your code your variable x_roots is a column vector, so your for loop iterates one time, and in that iteration i is equal to x_roots.

clear all
close all
syms n

w = @(x) 0.01.*x.^4-0.04.*x.^3+0.02.*x.^2+0.04.*x-0.03;
w_vec = [0.01 0.04 0.02 0.04 0.03]

w_vec = 1×5

0.0100 0.0400 0.0200 0.0400 0.0300

x = -2:.2:4;
w_y = w(x);

x_roots = roots(w_vec) % x_roots is a column vector ...

x_roots = 
  -3.6921 + 0.0000i
   0.2003 + 1.0521i
   0.2003 - 1.0521i
  -0.7084 + 0.0000i

count = 0;
for i=x_roots % ... so the for loop iterates once
    count = count+1;
    disp("Iteration: " + count)
    disp("i = ")
    disp(i)
end

Iteration: 1

i = 

  -3.6921 + 0.0000i
   0.2003 + 1.0521i
   0.2003 - 1.0521i
  -0.7084 + 0.0000i

But you want to iterate over each element of x_roots, not over each column of x_roots. An easy way to do that is to tranpose x_roots in the for expression so that what the for sees is a row vector, in which case it iterates once for each element:

x_realroots = [];

for i=x_roots.' % x_roots.' is a row vector, so the loop iterates multiple times
    ~round(imag(i))
    if (~round(imag(i)))
        round(imag(i))
        x_realroots = [x_realroots i];
    else
        continue;
    end
end

ans = logical
   1

ans = 0

ans = logical
   0

ans = logical
   0

ans = logical
   1

ans = 0

x_realroots

x_realroots = 1×2

-3.6921 -0.7084

A more general solution, which would work whether x_roots is a row vector or a column vector (or indeed an array of any size or dimensionality), would be, again in the for expression, convert x_roots into a column vector using (:) notation and then transpose to produce a row vector of all the elements of x_roots:

for i = x_roots(:).' % iterates over the elements of x_roots, regardless of its size
    % ...
end

[Konrad Suchodolski]

Thank you very much :D

[Voss]

You're welcome!

Answer 2

[madhan ravi]

 x_realroots = x_roots(abs(imag(x_roots)) < 1e-4) % 1e-4 tolerance

Answer 3

[Les Beckham]

Array indices in Matlab must be positive integers (or logical). You are trying to us xroots, a vector of complex numbers as the index in your for loop. Loop using integers instead and use that index to select each element of x_roots.

clear all
close all
syms n

w = @(x) 0.01.*x.^4-0.04.*x.^3+0.02.*x.^2+0.04.*x-0.03;
w_vec = [0.01 0.04 0.02 0.04 0.03];

x = -2:.2:4;
w_y = w(x);

x_roots = roots(w_vec)

x_roots = 
  -3.6921 + 0.0000i
   0.2003 + 1.0521i
   0.2003 - 1.0521i
  -0.7084 + 0.0000i

x_realroots = [];

% for i=x_roots
for i = 1:numel(x_roots)
    % ~round(imag(i))
    % if (~round(imag(i)))
    if (imag(x_roots(i)) == 0)
        % round(imag(i))
        % x_realroots = [x_realroots i];
        x_realroots = [x_realroots real(x_roots(i))];
    else
        continue;
    end
end
x_realroots

x_realroots = 1×2

-3.6921 -0.7084

x_realroots2 = real(x_roots(imag(x_roots) == 0))

x_realroots2 = 2×1

-3.6921 -0.7084

Note that you don't even need a loop.

Also, if you are just getting started with Matlab, I would highly recommend that you take a couple of hours to go through the free online tutorial: Matlab Onramp

[Konrad Suchodolski]

okay, but i was trying to use for to iterate elements of x_roots, not indexes.

for v = [1 5 8 17]
   disp(v)
end

something like this, but with x_roots

[Dyuman Joshi]

"You are trying to us xroots, a vector of complex numbers as the index in your for loop."

@Les Beckham, For loop index values can be anything. Do not confuse them with array indices.

And OP is not trying to use the for loop indices as array indices.

[Les Beckham]

Good point. I had misinterpreted what they were trying to do. Of course, the real "Matlab way" is to not have a loop in the first place.

w_vec = [0.01 0.04 0.02 0.04 0.03];
x_roots = roots(w_vec)

x_roots = 
  -3.6921 + 0.0000i
   0.2003 + 1.0521i
   0.2003 - 1.0521i
  -0.7084 + 0.0000i

x_realroots2 = real(x_roots(imag(x_roots) == 0))

x_realroots2 = 2×1

-3.6921 -0.7084