Loading 2D array files to Fortran variables

mrz1982

Hi,

I have created data files by compiling and running a c++ code. Now I
need to open and store the contents of these files in Fortran 77
variables. But I am not able to do this. The read command does not
return the values stored in the file.
The files are having some 2003 lines and 53 colomns. The values are in
the form "2.885531e+002" (positive) and "-7.687902e+001" (negative)
with 1 space seperating different values.

Thank you in advance!
/Z

jtbell

Can you show us the code that you tried to use for reading the file?

mrz1982

The entire code is very long, so I will only write the part where the file is opened and read. I have tried many different formats, here is the last format I tried with:

REAL EVg(2003,53)
open(1,file=''test.txt'')
read(1,10,iostat=k) EVg
10 format(53(G7E3,1X))
close(1)

the iostat has returned different values (-1, 0, 115, etc.) for different format cases I have tried. For the case above it returns 0, but the "EVg" array does not contain the array values.

jtbell

It's been a long time since I had to deal with this sort of thing, but as I recall, the basic principles are these: (1) When the number of values to be input exceeds the number of specifiers in the format statement, the format simply repeats itself. (2) All data in a single input statement is assumed to be on one line, unless the format specifies otherwise.

So I think your program is trying to read the entire array from the first line of the input file. Try adding a slash at the end of the format, which should skip to the next line:

10 format(53(G7E3,1X)/)

This may still give you problems if the file doesn't have a blank space after the last number on each line.

Also, I don't remember a format specifier like "G7E3", but my Fortran experience ended with Fortran 77.

mathmate

I don't either.
In fact, the enhanced format specifiers (E,F,G) are mostly for displaying pretty output.
For input, an E,F, or G specifier will accept both 123456.78 or 1.1234E6, as shown in the following example code. The input I gave was 1.123456E+012 1.123456e+013 .
The output was printed correctly.

Code:

C test fortran format
      REAL*8 A,B
      READ(5,1)A,B
C   1 FORMAT(2(F13.6,1X))
C   1 FORMAT(2(E13.6,1X))
    1 FORMAT(2(G13.6,1X))
      PRINT *,A,B
      STOP
      END

Do not forget the extra space at the end of the line, it is required.
If you still have problems, post a few lines of your input, using the code option so we see exactly how the data are laid out.
Also, check if the spaces between the data are aligned. The format above expects them to be aligned. If you have a problem with data alignment, write a c-program to read in and write them out aligned. That would be easier to do to figure out how to read it using free-format input.

mrz1982

Thank you for your response!
I should maybe add that I am working in Fortran 77 enviroment (due to limitations).
The data in the files basically look like:

Code:

+0.000000e+000 -7.281896e+002 -1.791944e+003 
+0.000000e+000 -7.856508e+002 -1.824643e+003 
-1.325604e+005 -8.411246e+002 -1.856220e+003

this is taken from somewhere in the middle of the file, at another place it looks like:

Code:

-1.045771e+005 +1.047517e+004 -1.180211e+003 
-5.662817e+004 +1.152035e+005 -1.173054e+003 
+1.187847e+004 +2.266384e+005 -1.165909e+003

as you can see the data are (now) aligned, (as I added '+' in front of positive values) and the extra space at each line end is there. There is also an line termination after the extra space of the last line.
The file reading procedure:

Code:

REAL EVg(2003,53)
open(1,file=''test.txt'')
read(1,10,iostat=k) EVg
10 format(53(E13.6,1X)/)
close(1)

returns k=0 and EVg a matrix of zeros.

mathmate

Indeed, an error in the format will cause the numbers read in to become zeroes.

Using the following code:

Code:

C testREAD fortran format
      REAL*8 EVg(6,3)
      open(1,file='testread.dat')
      DO 5 I=1,6
    5 READ(1,FMT='(3E15.6)',END=15)(EVg(I,J),J=1,3)
   15 close(1)
      DO 20 I=1,6
   20 WRITE(6,FMT='(SP,3(1PE15.6E3))')(EVg(I,J),J=1,3)
      END

and reading the six lines you supplied from a data file, I obtain the following output:

Code:

+0.000000E+000 -7.281896E+002 -1.791944E+003
+0.000000E+000 -7.856508E+002 -1.824643E+003
-1.325604E+005 -8.411246E+002 -1.856220E+003
-1.045771E+005 +1.047517E+004 -1.180211E+003
-5.662817E+004 +1.152035E+005 -1.173054E+003
+1.187847E+004 +2.266384E+005 -1.165909E+003

Hope that helps to solve your problem.

mrz1982

I have now tried with the following (on a somewhat smaller file [2001][12]),

Code:

open(1,file="test.txt")
DO t=1,2001
READ(1,'(12E14.6)',iostat=ios)(EVg(t,x),x=1,12)
ENDDO
close(1)
DO t=1,2001
WRITE(6,FMT='(SP,12(1PE15.6E3))')(EVg(5,x),x=1,12)
ENDDO

The iostat returns -1, if I change the format to

Code:

'12(E14.6,1X)'

as it should be iostat returns 100.
For both formats I still get a EVg to be an array of 0's.

Can there be a problem with the compiler I am using? I am using the "mingw32-make.exe" which runs a "g77 -c -O2 fortran.f" and a "g77 -s -o file.exe dimdef.o newmods.o comtac.o fgnmod.o usernl.o analyt.o devt69.o usrfun.o hopcod.o user10.o tpbig.a dislin.a -luser32 -lgdi32 -lcomdlg32".

mathmate

I cannot pronounce anything until you show the actual data you are using with 12 numbers on the same line. In Fortran, the size of each field is important. Even an extra space at the end of the line could be significant. The number of lines is not important in this case.

If you could post 4-5 lines of data, then I'll see how the data fits in with the format.
If we don't find anything there, then we may have to worry about the compiler.

mrz1982

Here are 5 lines taken from the middle of the data file:

Code:

+1.670000e-006 +0.000000e+000 -8.540156e+006 -1.585399e+003 -2.243492e+003 -2.763622e+003 -3.142946e+003 -3.299538e+003 -3.304998e+003 -3.342794e+003 -3.706140e+003 -4.368216e+003 
+1.680000e-006 -9.375357e+004 -8.037304e+006 -1.612638e+003 -2.255844e+003 -2.763778e+003 -3.130649e+003 -3.279551e+003 -3.282095e+003 -3.319731e+003 -3.677900e+003 -4.360687e+003 
+1.690000e-006 -2.859594e+005 -7.588826e+006 -1.639665e+003 -2.269052e+003 -2.765124e+003 -3.119963e+003 -3.261429e+003 -3.261483e+003 -3.300265e+003 -3.655372e+003 -4.347075e+003 
+1.700000e-006 -6.289417e+005 -7.239074e+006 -1.666989e+003 -2.283188e+003 -2.767609e+003 -3.110709e+003 -3.244538e+003 -3.242276e+003 -3.283470e+003 -3.638555e+003 -4.326224e+003 
+1.710000e-006 -1.137889e+006 -7.009578e+006 -1.694831e+003 -2.298037e+003 -2.770898e+003 -3.102424e+003 -3.228056e+003 -3.223421e+003 -3.268063e+003 -3.626592e+003 -4.298185e+003

It is a space after the last values of every line (the option apperently removes extra space after the last value).

mathmate

Sorry it took a little more time than I thought would take, but I have remodelled the program to make it a little more general in order to read in a variable number of lines and fields. It seems to read in and write correctly.
You could try with this new version and make adaptations required for your purposes:

Code:

C testREAD fortran format
      REAL*8 EVg(5,12)
      CHARACTER *20 READFMT
      CHARACTER *25 WRITEFMT
      NLINES=5
      NFIELDS=12
      WRITE(READFMT,'(1H(I2,5HE15.61H))')NFIELDS
      WRITE(WRITEFMT,'(4H(SP,,I2,12H(1PE15.6E3)) )')NFIELDS
      PRINT *,READFMT,WRITEFMT
C     (12E15.6)           (SP,12(1PE15.6E3))
      OPEN(1,FILE='testread.512',FORM='FORMATTED',IOSTAT=IOS,RECL=181)
      IF(IOS.NE.0) THEN
        PRINT *,'ERROR IOS=',IOS
        STOP
      ENDIF
      DO 5 I=1,NLINES
      READ(1,FMT=READFMT,END=15)(EVg(I,J),J=1,NFIELDS)
    5 CONTINUE
   15 close(1)
      DO 20 I=1,NLINES
      WRITE(6,FMT=WRITEFMT)(EVg(I,J),J=1,NFIELDS)
   20 CONTINUE
C     FORMAT(SP,12(1PE15.6E3))
      STOP
      END

Output:

Code:

+1.670000E-006 +0.000000E+000 -8.540156E+006 -1.585399E+003 -2.243492E+003 -2.763622E+003 -3.142946E+003 -3.299538E+003 -3.304998E+003 -3.342794E+003 -3.706140E+003 -4.368216E+003
+1.680000E-006 -9.375357E+004 -8.037304E+006 -1.612638E+003 -2.255844E+003 -2.763778E+003 -3.130649E+003 -3.279551E+003 -3.282095E+003 -3.319731E+003 -3.677900E+003 -4.360687E+003
+1.690000E-006 -2.859594E+005 -7.588826E+006 -1.639665E+003 -2.269052E+003 -2.765124E+003 -3.119963E+003 -3.261429E+003 -3.261483E+003 -3.300265E+003 -3.655372E+003 -4.347075E+003
+1.700000E-006 -6.289417E+005 -7.239074E+006 -1.666989E+003 -2.283188E+003 -2.767609E+003 -3.110709E+003 -3.244538E+003 -3.242276E+003 -3.283470E+003 -3.638555E+003 -4.326224E+003
+1.710000E-006 -1.137889E+006 -7.009578E+006 -1.694831E+003 -2.298037E+003 -2.770898E+003 -3.102424E+003 -3.228056E+003 -3.223421E+003 -3.268063E+003 -3.626592E+003 -4.298185E+003

mrz1982

Thank you, it is working now. I was stuck with these lines for more than 2 weeks.... but now everything is working.

mathmate

Glad that everything works out, come and visit us often!

Similar Threads for: Loading 2D array files to Fortran variables
Thread	Forum	Replies
array shuffling in fortran	Programming & Comp Sci	1
Fortran 90 (need to read in binary data files into arrays)	Math & Science Software	1
Accessing Fortran Modules within a Fortran library from Fortran	Programming & Comp Sci	0
Fortran 90 question about reading files with text	Programming & Comp Sci	1
Implicit Variables in FORTRAN	Programming & Comp Sci	2

jtbell Mentor	Can you show us the code that you tried to use for reading the file?

mrz1982	Thank you, it is working now. I was stuck with these lines for more than 2 weeks.... but now everything is working.

mathmate	Glad that everything works out, come and visit us often!