forums.silverfrost.com

christyleomin · Joined: 08 Apr 2011 Posts: 155

Hi,

I am relatively new to Fortran and havce a piece of code from smeone which I am not abe to fully follow;

dimension x(3,*),y(3,*)
character*70 check

undef = -1.0d30
icount = 0
do i = 1 , nav
if(scr(3,i) .ne. undef) then
icount = icount + 1
x(1,icount) = y(1,i)
x(2,icount) = y(2,i)
x(3,icount) = y(3,i)
endif

if(x(3,i).eq.undef)then

call something()
endif

enddo

Please can anyone help?

What is undef here? Is it a FORTRAN keyword?

Please help--

PaulLaidler · Posted: Fri Jul 01, 2011 7:20 am Post subject:

It is a REAL variable. Its type is governed by the Fortran standard relating to "implicit" typing.

davidb · Joined: 17 Jul 2009 Posts: 560 Location: UK

I can't follow it either, but for different reasons. As Paul says undef is just a real variable in this code. However, you should not be assigning it the
value of a double precision constant (-1.0d-30). Moreover, x(3,i) which you use in the test is not defined, e.g. if the first test gives .false. it is not defined with i = 1.

Your code basically copies the elements of y where scr(1:3,:) is not equal to undef to x(1:3,:).

Here is my version of the copy part.