A problem with file reading

Help
1

Casus: a file is being read, looking like so:
*
key
new line
new line
*
key
new line
new line
*
So relevant text blocks start and finish with * The amount of lines per block varies. The data item searching for the appropriate keys of the data to be displayed is “seek”

My question is: could the coding be improved?

Patrick

10 do while string(1:1).NE.‘*’)
read(1,“(a)”,iostat=stat)string
end do
read(1,“(a)”,iostat=stat)string
if (index(string,seek(1:3)).NE.1) goto 10

do while string(1:1).NE.'') . . . . . the astrisk between the quotes is not being shown by the site
read(1,“(a)”,iostat=stat)string
print_, string !the program uses a subroutine to print
end do

2

This code was generated using ChatGPT. Does it meet your expectations?

program ReadFile

   implicit none

   character(len=5)   :: keyword = "key1"
   character(len=100) :: line
   logical            :: found_keyword = .false.
   integer            :: nunit
   integer            :: status

   ! Open the file for reading
   open(newunit=nunit, file='testfile.txt', status='old', action='read')

   ! Read the file until the end or until keyword is found
   do
       read(nunit, '(A)', iostat=status) line
       if (status < 0) exit ! End of file reached
       if (trim(line) == keyword) then
           found_keyword = .true.
           exit
       endif
   end do

   ! Print the text block for keyword
   if (found_keyword) then
       do
           read(nunit, '(A)', iostat=status) line
           if (status < 0 .or. line(1:1) == '*') exit ! End of file or next key reached
           write(*, '(A)') line
       end do
   else
       write(*, '(A)') 'keyword not found.'
   endif

   ! Close the file
   close(nunit)

end program ReadFile

testfile.txt:

*
key1
Lorem ipsum dolor sit amet, consectetur adipiscing elit.
Sed ac elit vitae purus aliquam luctus.
Aliquam vitae fringilla risus.
Proin interdum dignissim diam at mattis.
Sed sed metus tristique, tincidunt sapien a, consectetur felis.
Mauris eu dolor a mauris tincidunt elementum.
Vivamus ac aliquam massa.
Donec venenatis aliquet velit, eu scelerisque justo luctus ac.
Vestibulum ante ipsum primis in faucibus orci luctus et ultrices posuere cubilia Curae;
Nam ullamcorper est non mi cursus, sed vestibulum ante pellentesque.
*
key2
Proin interdum dignissim diam at mattis.
Sed sed metus tristique, tincidunt sapien a, consectetur felis.
Donec venenatis aliquet velit, eu scelerisque justo luctus ac.
Vivamus ac aliquam massa.
*
key3
Sed sed metus tristique, tincidunt sapien a, consectetur felis.
*
key4
Lorem ipsum dolor sit amet, consectetur adipiscing elit.
Vestibulum ante ipsum primis in faucibus orci luctus et ultrices posuere cubilia Curae;
Nam ullamcorper est non mi cursus, sed vestibulum ante pellentesque.
*
3

Can you guarantee that “key” and “new line” don’t start with a “*” ?

Given this line is “*”, you need to:

  1. Read nexl line, as a Key, checking if it is what you want
  2. Read lines to identify end of block, ie until “*” is reached.
  3. repeat the process, as you have the start of the next block.

My logic would be to: read_next_line then respond, depending if you are in stages 1, 2 or 3

4

Thank you for the code ! ! !

          Got it to work :

      -  if (trim(line) == keyword) then . . .
      + if (trim(line(1:3)) == keyword(1:3)) then . . .

        "instring?" would have been better

I was unable to respond sooner, for which sorry.
Patrick.

1 Like
5

@Ali
Thanks for the example of using ChatGPT !

Is this what ChatGPT generated, or did you tidy it up ?

It is interesting that the code is not correct, as my understanding of the thread is a keyword can only follow a terminator line “*”

Not sure about the single * terminator selection either, as **** might have been a stronger terminator ?

6

The code was generated entirely using GPT, and I didn’t expect it to compile correctly without any errors. Therefore, I hastily posted it here without making any changes (except for adding ‘nunit’). I thought GPT could provide a quick idea.

7

That is certainly the case. It will be interesting to see where this approach can take us.