[fortran90] have problem in real array index

[s_hy]

hi all...i have problem in understanding in array index. my codes are as below

subroutine  tmz2d
implicit none


double precision                   :: phi
double precision                   :: cosphi,sinphi
double precision                   :: d,dp,dpp,d1,d2
integer                            :: i,j,m,h
integer                            :: minit,mlast
integer                            :: l,p
integer                            :: iinit,ilast,hlast,pinit,plast
integer                            :: n
integer                            :: linit,llast
integer                            :: jinit,jlast
double precision,dimension(200)    :: Hm,Em
double precision                   :: Einc,Hinc
double precision,dimension(200,200):: Ezinc,Hxinc,Hyinc
double precision, parameter        :: pi = 3.14159265
character(len=20)                  :: filename
 

 minit = 1
 mlast = 100
 linit = 1
 llast = 100
 pinit = 1
 plast = 100

 iinit = 1 
 ilast = 100
 jinit = 1
 jlast = 100


 phi = pi/4.0
 cosphi = cos(phi)
 sinphi = sin(phi)
 print *, 'cosphi=',cosphi,'sinphi=',sinphi


do n = 1,100
   
  write (filename, "('data',I3.3,'.dat')") n
  open (unit=130,file=filename)


 call bg1d
 !print *, 'see me2'

  do m = 2,2*ilast-1
    do p = 2,2*jlast-1
      if (Mod(m,2)/=0) then
	if (Mod(p,2)/=0) then
         d = cosphi*(m - minit) + sinphi*(p - pinit)
         dp = d - int(d)
         Einc = (1-dp)*Em(int(d))+dp*(Em(int(d)+1))
         Ezinc(m,p+2) = Einc
         write(130,*) Ezinc(m,p+2)
        end if
      end if
    end do
  end do

  do m = 2,2*ilast-1
    do p = 2,2*jlast-1
      if (Mod(m,2) == 0) then
	if (Mod(p,2) == 0) then
         d = cosphi*(m - minit) + sinphi*(p - pinit)
         dpp = d+1
         dp = dpp-int(d)
           Hinc = (1-dp)*Hm(int(dpp)-0.5)+dp*Hm(0.5+int(dpp))
           Hxinc(m,p+1) = sinphi*Hinc
           Hyinc(m-1,p+2) = -cosphi*Hinc
        end if
      end if
    end do
  end do

end do !n
print *, 'see me3'


end subroutine tmz2d

and i got this warning when execute

subroutines.f90:167.48:

Hinc = (1-dp)*Hm(int(dpp)-0.5)+dp*Hm(0.5+int(dpp))
1
Warning: Extension: REAL array index at (1)
subroutines.f90:167.28:

Hinc = (1-dp)*Hm(int(dpp)-0.5)+dp*Hm(0.5+int(dpp))
1
Warning: Extension: REAL array index at (1)


###########################################

the problem come from line 167 and i found out that fortran do not allow real index...(int(dpp) -0.5)..can anyone give me an advice how to solve this problem.

thank you in advance

 

[Borek]

The obvious answer (which should remove the warning, but doesn't have to be correct for your problem) is to use int(dpp-0.5) - right now you are converting dpp to an int, then subtracting (or adding) 0.5 converting it back to a real. Do the subtraction first, conversion to an int later.

 

[gsal]

O.k., you found out that Fortran does not allow "real" index; more precisely, it allows a "real argument" but the final index needs to be integer...to that end, it allows a "real argument" which gets automatically typecast into an integer. If "real argument" wasn't allowed, you would get a compiler "error" instead of a "warning".

Say, is there a chest of drawers in your room? Does it have drawer 1.37?...It's the same with arrays.

 

[s_hy]

thanks Borek, its worked and removed the warning

thanks gsal for the explanation..i am really new in fortran90

 

[LudusRex]

Hello,

It seems like the warning you are receiving is due to using a real number as an index for your array. In Fortran, array indices must be integers. To solve this problem, you can convert your real number to an integer using the INT function.

For example, instead of using int(dpp)-0.5, you can use int(dpp-0.5). This will convert the result of dpp-0.5 to an integer before using it as an index.

You can also use the FLOOR function which will round down the real number to the nearest integer. So instead of int(dpp)-0.5, you can use floor(dpp).

I hope this helps. Happy coding!