About choosing negative and positive elements from vector of real, imaginary and complex numbers.

Hi here is my problem and thanks in advance for reading.

I solve an equation (eqn) that returns eight roots, they will be a mix of real, imaginary, and complex. I want to separate them into two vectors such that 'vector 1' contains the roots that are positive real, positive imaginary, and if complex - the complex number which has real positive part (irrespective of the sign on the imaginary part). The remaining roots will be in the second vector 'vector 2'. Thus each vector will usually contain four elements.

I have tried this so far,

vector1 = solve(eqn);

vector1(real(vector1)<0)=[ ]; % removes negative real roots

vector1(imag(vector1)<0)=[ ]; % removes negative imaginary roots

The problem is that the command to remove the imaginary roots also removes the complex roots with positive real part and negative imaginary part i.e. a-bi, but I want to keep those. I have thought about using an 'if' statement to say that if the number is complex then only remove the complex numbers with negative real part, but I haven't managed it.

Any advice would be awesome.

EDIT:

I managed to do it like this

eqn1=solve(eqn);

zimpos=[]; % to add the positive imaginary values

for k = [1:length(eqn1)]

    if real(eqn1(k))==0;

       zimpos=[zimpos,(imag(eqn1(k)))];

       zimpos(zimpos>0)=[];

    end

end

    zimpos=zimpos*i;

zrealpos=[]; % to add the complex or real numbers with positive real part

for k= [1:length(eqn1)]

    if real(eqn1(k))>0;

        zrealpos=[zrealpos,(eqn1(k))];

    end

end

zpos=[zrealpos zimpos]; % the final vector.

If anyone has any suggestions on improving this I would be very grateful.

Thanks.

Answer 1

[Jan]

eqn1 = solve(eqn);

hasImag = any(imag(eqn1(:)));                % ~ISREAL might be wrong
m1      = ~hasImag & real(eqn1) > 0;         % positive real
m2      = real(eqn1) == 0 & imag(eqn1) > 0;  % positive imaginary
m3      = hasImag & real(eqn1) > 0;          % complex and real positive part
m       = m1 | m2 | m3;

vector1 = eqn1(m);
vector2 = eqn1(~m);

[AllKindsofMath AllKinds]

Worked like a charm thank you very much. The only edit I made was that I needed to sort the roots so instead of the last two lines I did

vector1 = sort(eqn1(m)); vector2 = sort(eqn1(m))*-1;

Again thanks.