Who can define a complex matrix in MATLAB? Pleale translate the FORTRAN into Matlab code?

In FORTRAN the function COMPLEX can define a complex matrix ,like as COMPLEX B(0:20,0:10). But in MATLAB, there is no COMPLEX function. I have a FORTRAN code. Can you translate it into Matlab code?

PROGRAM MAIN
    DOUBLEPRECISION A(0:20,0:10)             
        COMPLEX  B(0:20,0:10) 
     DO I=0,20
            DO J=0,10
                   A(I,J)=2*I+J
               B(I,J)=CMPLX(A(I,J),0.0)
            ENDDO
          ENDDO 

       ! ! !  Output calculation results: A and B   ! ! !
          OPEN(1,FILE='A.DAT')
          WRITE(1,10) ((A(I,J),J=0,10),I=0,20)
      CLOSE(1)
          OPEN(2,FILE='B.DAT')
          WRITE(2,20) ((B(I,J),J=0,10),I=0,20)
      CLOSE(2)   
  10  FORMAT(X,11F14.8)   
  20 format( 22f14.8 )                               
    END

[John D'Errico]

It is time to learn MATLAB. Read the getting started tutorials.

[per isakson]

"But in MATLAB, there is no COMPLEX function." See Complex Numbers and complex, Create complex array

Answer 1

[KSSV]

In matlab default class is double.....you need not to initialize as you do in FORTRAN. If you want to define a complex number, you do the following:

a = 5 ; 
b = 6 ;
z = a+1i*b ;   % complex number 
real(z)      % gives real part
imag(z)      % gives complex part

Same is the case with arrays.

[per isakson]

Matlab

for I = 1 : 21

[Paozhuan Zhang]

Thank you very much. According to your program method ,the MATLAB code as:

B Is a real matrix with 21 X 11. But in fortran B is a comple xmatrix. SO I cannot obtain the same result as the FORTRAN.

[per isakson]

"same result" I guess that depends on how you define "result". See See the comment by John D'Errico

>> B = 1 + 0i;
>> class(B)
ans =
double
>> imag(B)
ans =
     0

and

>> b = complex( 1, 0 )
b =
   1.0000 + 0.0000i
>> class(b)
ans =
double

>> b1 = 1;
>> imag(b1)
ans =
     0

[Walter Roberson]

In MATLAB, if all of the complex entries in the matrix are 0, then MATLAB removes the complex portion, leaving it real-only valued. However, in nearly all cases MATLAB treates real-only valued matrices as being equivalent to a complex matrix whose complex parts all read out as zero. In the above code, real(B) would extract the real component and imag(B) would extract the imaginary component which would come out all zero if B happened to be a real-only value.

Answer 2

[James Tursa]

A = bsxfun(@plus,2*(0:20)',(0:10));
B = complex(A);

Same resulting matrices as Fortran, except that MATLAB indexing is always 1-based. I.e., the imaginary part of B is physically present in memory, same as Fortran, even though the imaginary values are identically zero.