Writing specific parts of the file to another file - fortran

Question

I am writing a code to read huge file (File XYZ) which has a couple of objects (say A,B,C,D,E... there might be more and each object has hundreds of data files). Now I need to read the file into Fortran and select an object for example 'object B' and write all the information of 'object B' into a new file which I create using the fortran code.
Each object (from A-....) had information consisting of header files and an array of data (a matrix(i,j) where i varies for each file and j=4 ). I have to read in both the header and array data into the new file and these objects (A-....) are not ordered sequentially (suppose after A may be C or D).
I have written a small code but it doesn't seem to work well as I am either able to output only the header but not the array or nothing at all attaching a data file for an object and code.
program trail
implicit none
CHARACTER (LEN=5) :: thumb='3C286', SOURCE
INTEGER :: pair(4), n,io, allocstat
real, dimension(:,:),allocatable:: a
CHARACTER (LEN=20) :: POLCODE, DEC, RA
CHARACTER (LEN=40) :: TELESCOPE, NAME_MY, DATA_FILENAME, object , date, TEXT, HEADER, restfreq,BANDWIDT, TSYS, OBSTIME, ANGLE
character(len=180) :: Outdat
character(len=75) :: ra_1, residu, flux, vlafit
open(2, FILE='trail9.txt',POSITION='APPEND')
do
open(1, File="column_3C286_iteration1.txt",status='old' )
read(1,*), SOURCE
10 if(Source.eq.'OJ287') then
! READING THE DATA
read(1,'(A40)') DATA_FILENAME
read(1,'(A40)') TELESCOPE
read(1,'(A40)') NAME_MY
read(1,'(A40)') data_filename
read(1,'(A40)') OBJECT
read(1,'(A40)') restfreq
read(1,'(A20)') polcode
read(1,'(A40)') angle
read(1,'(A40)') date
read(1,'(A40)') bandwidt
read(1,'(A40)') tsys
read(1,'(A40)') obstime
read(1,'(A20)') ra
read(1,'(A20)') dec
read(1,'(A40)') text
read(1,'(A40)') header
read(1,'(A180)') OUTDAT
read(1,"(T2, A3,T30,A4, T50,A6,T67,A6)"),ra_1, residu, flux, vlafit
n=0
do
read(1,*,iostat=io) pair
if (io/=0) exit
n = n + 1
end do
allocate(a(n,4),stat=allocstat)
do i=1,n
read(1,*) (a(i,j),j=1,4 )
enddo
if (index(data_filename,'#')>0 ) then
!WRITING THE DATA
write(2,'(A40)') DATA_FILENAME
write(2,'(A40)') TELESCOPE
write(2,'(A40)') NAME_MY
write(2,'(A40)') data_filename
write(2,'(A40)') OBJECT
write(2,'(A40)') restfreq
write(2,'(A20)') polcode
write(2,'(A40)') angle
write(2,'(A40)') date
write(2,'(A40)') bandwidt
write(2,'(A40)') tsys
write(2,'(A40)') obstime
write(2,'(A20)') ra
write(2,'(A20)') dec
write(2,'(A40)') text
write(2,'(A40)') header
WRITE(2,'(A180)') OUTDAT
Write(2,"(T2, A3,T30,A4, T50,A6,T67,A6)"),ra_1, residu, flux, vlafit
do i=1,n
write(2,"(T1,F10.4: T27,F10.4,T47, F10.4, T63,F10.4, T75,F10.4 )"), (a(i,j), j=1,4)
end do
end do
end if
end do
end program trail

Answer 1

Please try changing the type of pair from integer to real and modify the code as
real :: pair(4)
integer :: i
n = 0
do
read(1,*,iostat=io) pair
if (io/=0) exit
n = n + 1
end do
do i = 1, n + 1
backspace( 1 )
enddo
In this way, the read statement counts the lines including four floating-point numbers and stops reading when non-numeric characters (such as #) are found in the line. Please note also that backspace() is performed n+1 times so that the file pointer goes back to the top of the current data section.
Although not very sure, I also think that the following open statement
open(1, File="column_3C286_iteration1.txt",status='old' )
should be written before the outermost DO loop if this corresponds to the data file.
To summarize, I suggest the following modification:
real :: pair(4)
integer :: i
open(1, File="column_3C286_iteration1.txt",status='old' )
do
read(1,*,end=5000), SOURCE
if( Source .eq.'OJ287' ) then
! read the remaining header and data lines in one block.
read(1,'(A40)') DATA_FILENAME
read(1,'(A40)') TELESCOPE
...
read(1,"(T2, A3,T30,A4, T50,A6,T67,A6)"),ra_1, residu, flux, vlafit
n = 0
do
read(1,*,iostat=io) pair
if (io/=0) exit
n = n + 1
end do
do i = 1, n + 1
backspace( 1 )
enddo
if ( allocated( a ) ) deallocate( a )
allocate( a( n, 4 ) )
do i = 1, n
read( 1, * ) a( i, 1:4 )
enddo
!! write out the obtained data to file 2.
endif
enddo
5000 continue
Edit
If we simply want to copy one block of "header + data" to a different file, it is more straight-forward to echo all the lines in the block without formatting. For example, the program may look like this.
character(200) :: str, field
integer :: ios
real :: dum
open( 10, file="input.dat", status="old" )
do
read( 10, "(a)", end=5000 ) str
if ( trim(str) == "OJ287" ) then !! search for the target block
print "(a)", trim(str) !! print the tag
!! read one block.
do
read( 10, "(a)", end=5000 ) str
read( str, * ) field !! get the first field of a line
if ( field(1:1) == "#" .or. field(1:2) == "ra" ) then
print "(a)", trim(str) !! header part
else
read( str, *, iostat=ios ) dum !! test if this line starts with a number
if ( ios == 0 ) then
print "(a)", trim(str) !! data part
else
backspace( 10 )
exit
endif
endif
enddo
endif
enddo
5000 continue