Question

This is somewhat related to my recent post about Fortran stream and the like: Converting data stored in Fortran 90 binaries to human readable format.

I am attempting to write a simple array of integers to a file, and then use Fortran’s READ function to then read in the binary that I have created. I am using stream by including ACCESS="STREAM" in my OPEN directive. I have the following code:

MODULE streamtest2subs

  IMPLICIT NONE

CONTAINS

SUBROUTINE writeUstream(myarray)
  IMPLICIT NONE
  INTEGER, INTENT(IN), DIMENSION(4,10) :: myarray
  INTEGER :: myvalue = 12345, mypos
  OPEN(UNIT=11, FILE="ustream.demo", STATUS="REPLACE", ACCESS="STREAM")
  WRITE(11) myarray
  CLOSE(UNIT=11)
END SUBROUTINE writeUstream

SUBROUTINE readUstream
  IMPLICIT NONE
  INTEGER :: test1, test2, test3
  INTEGER :: n
  OPEN(UNIT=42, FILE="ustream.demo", STATUS="OLD", ACCESS="STREAM")
  READ(42, POS=1) test1
  READ(42, POS=2) test2
  READ(42, POS=3) test3

  WRITE(*,*) "This is the output:"
  WRITE(*,*) test1
  WRITE(*,*) test2
  WRITE(*,*) test3
END SUBROUTINE readUstream

END MODULE streamtest2subs

PROGRAM streamtest2
  USE streamtest2subs
  IMPLICIT NONE
  INTEGER :: i, j, k
  INTEGER, DIMENSION(4,10) :: a

  WRITE(*,*) "This is my input array:"
  k=1
  DO i=1,4
    DO j=1,10
      a(i,j)=k
      WRITE(*, "(i3)", ADVANCE="NO") a(i,j)
      k=k+1
    END DO
    WRITE(*,*)
  END DO
  WRITE(*,*)

  CALL writeUstream(a)
  CALL readUstream
END PROGRAM streamtest2

However, when I compile this using gfortran and run it, I get the following output:

 This is my input array:
  1  2  3  4  5  6  7  8  9 10
 11 12 13 14 15 16 17 18 19 20
 21 22 23 24 25 26 27 28 29 30
 31 32 33 34 35 36 37 38 39 40

 This is the output:
           1
   184549376
      720896

Why is it that the output is so complex? Is it that READ is reading the ustream.demo file as a string instead of as an integer? However, when I change the type of test1, test2, and test3 to string, my output is simply a series of three blank lines.

Am I using the POS directive in READ incorrectly? I think that POS specifies the character number in the output (although I am not sure if the elements of the array are delimited in any way); is this correct?

Thank you very much for your time!

Answer 1

For the purposes that you have described in your previous question, I think that you are programming this more complicated than is necessary. Assuming that you want a binary file without extra record structure, you want unformatted and stream. You can read the file in the same way that you wrote it — you don’t need to use POS — unless the purpose is to learn how to use POS.

My tested example is based on a Fortran program in which I read a binary file that was written by someone else, almost certainly with a C program. The file consists of headers, followed by arrays of variable lengths. I open the file:

open ( unit=75, file=FileName, status='old', access='stream', form='unformatted', action='read' )

I read a header (a variable that is a user-define type with many sub-variables):

read (75) header

I allocate the array to the length that was read into one field of the header, then I read the array:

allocate ( array (1:header % ArrayLen) )
read (75) array

Then I process the data in the array.
Then I repeat until end of file (not shown in the code samples).

Very easy … no need to calculating position in the file and use the POS keyword of READ.