Namelist Error (Fortran runtime error: End of file)

Help
1

I want to read variables from the program. The main program is

PROGRAM nltest
  IMPLICIT NONE
  real        :: A
  REAL        :: B
  real        :: C, D
  
  namelist /FirstName/ A, B
  namelist /SecondName/ C, D
  
  OPEN (UNIT=10, FILE='FirstData.dat', STATUS='UNKNOWN')
  READ (10, NML=FirstName)
!  CLOSE (10)
 
!  OPEN (UNIT=10, FILE='FirstData.dat', STATUS='UNKNOWN')
  READ (10, NML=SecondName)
  CLOSE (10)
  
  PRINT 100, A, B
  
100 FORMAT (' A = ', f8.4, ' B = ', f8.4)
END PROGRAM nltest

The FirstData.dat file is

&FirstName
 A=  0.50    ,
 B=  0.02    ,
 /
&SecondName
 C= 0.40    ,
 D= 0.01    ,
 /

I am getting the error

gfortran main.f90 -o main

C:\Users\owner\Desktop>main
At line 15 of file main.f90 (unit = 10, file = 'FirstData.dat')
Fortran runtime error: End of file

Error termination. Backtrace:
#0  0x22a1fb85 in next_record_r
        at ../../../gcc-14.1.0/libgfortran/io/transfer.c:3899
#1  0x22a2c5a4 in finalize_transfer
        at ../../../gcc-14.1.0/libgfortran/io/transfer.c:4452
#2  0x229f17c5 in ???
#3  0x229f18b3 in ???
#4  0x229f139d in ???
#5  0x229f14d5 in ???
#6  0xf55f7373 in ???
#7  0xf6a7cc90 in ???
#8  0xffffffff in ???

So is it not possible to use multiple namelists? Or how to use different namelists from the file?

2

I’m using namelists in my codes and I can read several namelists in a single file without problem.

Your code compiles fine and when running with the FirstData.dat file, I don’t get any error. The output is:

A = 0.5000 B = 0.0200

As expected.

3

Is the error of the same origin as one discussed here? Fortran runtime error: End of file | ParaMonte: Parallel Monte Carlo and Machine Learning Library

3 Likes
4

You are probably missing a line terminator on the last line, and @shahmoradi points to a way to
avoid that error. Since a NAMELIST file skips lines not in a NAMELIST group definition instead of
adding blank lines to the end of the file I prefer a line of printable text such as “END-OF-FILE”.

NAMELIST group input file
Sat Jan 18 2025

&FirstName
 A=  0.50    ,
 B=  0.02    ,
 /

&SecondName
 C= 0.40    ,
 D= 0.01    , 
 /
END-OF-FILE

Adding printable text after the last set of data is recommended as it is less likely to be inadvertently removed. The problem is even if you add IOSTAT= to the READ statements and test for end-of-file you also get end-of-file when the NAMELIST group is not present or is truncated, so testing for an end-of-file status after the read is ambiguous. It is a useful feature of NAMELIST that it allows other text in the file, as it allows easy annotation of your data files so you can track their pedigree.

If you do “wc -c” on your file it probably shows 86 characters. One method to see if it is properly terminated is to do "cat -v -e -t $FILENAME" and if the last line does not end with a $ it does not have a line terminator.

For people that have cut and pasted the example input file you very likely get 87 characters. To make a “bad” file you could use something like

mv Firstdata.dat Firstdata.dat.good
head --bytes=-1  Firstdata.dat.good >Firstdata.dat

or if your head(1) command cannot handle that option

mv Firstdata.dat Firstdata.dat.good
dd iflag=count_bytes,binary if=Firstdata.dat.good count=86 of=Firstdata.dat

to make the file with the missing line terminator that will cause an error with some compilers.

so the best practice is not to end a NAMELIST file with the last line being a NAMELIST group terminator as noted above, but I recommend it be a printable text line not a blank line that you add to the end of the file, or a sacrificial NAMELIST definition that you do not read like

&EOF /
3 Likes
5

What is ch here? I get the error if it is there. Removing it does not generate any error. I am using Windows 10. Does it have to do with Linux?

6

has to do with my mouse jumping around because of an overly sensitive mousepad :>. It was a typo . I corrected it. :slight_smile:

1 Like