Printing a complex matrix Fortran - matrix

Question

The complex matrix is declared this way:
complex(8) :: matrix(:,:)
How can I print this matrix with each element as: (a, b) or a+ib, and in a nxn format? (by that I mean as a square matrix, with a row per line, so there will be n rows and n columns)
This is the way I would print a real matrix with the format I want:
do i=1,n
do j=1,n
write(*, fmt="(f0.2, tr2)", advance="no") matrix(i,j)
end do
write(*, fmt="(a)") " "
end do
But I'm not sure how to translate this to a complex matrix

Answer 1

How can I print this matrix with each element as: (a, b)
Supposing you already know that (a b) is the default printing fotmat for complex type, Why isn't this just enough?
do j=1,n
write(*, *) matrix(:,j)
end do
The output would be something like:
(10.000000000000000,-20.000000000000000) (10.000000000000000,-20.000000000000000) (10.000000000000000,-20.000000000000000)
(10.000000000000000, 20.000000000000000) (10.000000000000000, 20.000000000000000) (10.000000000000000, 20.000000000000000)
If you want something more customized, you could try something like this (adjusting the field width and precision):
do j=1,n
write(*, "(*('('sf6.2xspf6.2x'i)':x))") matrix(:,j)
end do
That produces something like this:
( 10.00 -20.00 i) ( 10.00 -20.00 i) ( 10.00 -20.00 i)
( 10.00 +20.00 i) ( 10.00 +20.00 i) ( 10.00 +20.00 i)

Answer 2

So far this is what has worked for me. Taking into account Clinton's advice:
character(19) fmt
fmt = '(F7.2,"+",F7.2,"i")'
do i=1,n
do j=1,n
fmt(8:8) = MERGE('+',' ',imag(a(i,j)).gt.0)
write(*,fmt, advance="no") a(i,j)
end do
write(*, fmt="(a)") " "
end do
And the output is:
-0.26 -0.00i -0.00 -0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i
0.00+ 0.00i -0.25 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i
0.00 0.00i 0.00 0.00i -0.05 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i
0.00 0.00i 0.00 0.00i 0.00 0.00i -0.05 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i
0.00+ 0.00i -0.46 -0.00i -0.00+ 0.00i -0.00+ 0.00i -0.50 0.00i 0.00 -0.00i 0.00 0.00i -0.00 0.00i
0.32+ 0.00i -0.00+ 0.00i 0.00 -0.00i -0.00 -0.00i 0.00+ 0.00i -0.27 0.00i 0.00 0.00i 0.00 0.00i
-0.00+ 0.00i 0.00 -0.00i -0.00 -0.00i 0.24+ 0.00i 0.00 0.00i 0.00 0.00i -0.18 0.00i 0.00 0.00i
0.00 0.00i -0.00 -0.00i 0.24+ 0.00i 0.00 -0.00i 0.00 0.00i 0.00 0.00i 0.00 0.00i -0.18 0.00i
If someone has a better suggestion I'll be glad to hear it :)

Answer 3

Here is something that has worked in the past, it could be fine tuned quite a bit
!! compile and link with gfortran -I/usr/include -o PrintComplex PrintComplex.f90
Program PrintComplex
use, intrinsic :: iso_c_binding
implicit none
integer, parameter :: N=16
integer :: k
real (kind=c_double) :: val(N)
complex (kind=c_double_complex) :: in(N)
character(19) fmt
fmt = '(F7.2,"+",F7.2,"i")'
val=(/(sin(3.14159d0*float(k)/3.d0),k=1,N)/)
in=cmplx(val,-val/2)
print *,"in"
do k=1,N
fmt(8:8) = MERGE('+',' ',imag(in(k)).gt.0)
write(*,fmt)in(k)
end do
End Program PrintComplex
The output is:
in
0.87 -0.43i
0.87 -0.43i
0.00 -0.00i
-0.87+ 0.43i
-0.87+ 0.43i
-0.00+ 0.00i
0.87 -0.43i
0.87 -0.43i
0.00 -0.00i
-0.87+ 0.43i
-0.87+ 0.43i
-0.00+ 0.00i
0.87 -0.43i
0.87 -0.43i
0.00 -0.00i
-0.87+ 0.43i