process input print file

[Anonymous]

Hi, I'm trying to read a (legacy dos) print file and output a CSV file to view the print file in Excel.
I'm processing the input byte by byte taking only printable characters and outputing comma seperated fields - enclosing text in quotes.
Processing byte by byte how do I check that I've arrived at the end of the print line - it terminates hex '0d0a'.?
TheFile Control is

::cobol
           SELECT PRN-IN
           ASSIGN TO 'C:\GCcobsrc\S041IBUD.PRN'
                          ORGANIZATION IS LINE SEQUENTIAL.

The File Section

::cobol
       FILE SECTION.
       FD  PRN-IN.
       01  PRN-LINE    PIC X(144).
       01  FILLER REDEFINES PRN-LINE.
           05  PRN-CHAR OCCURS 144 PIC X.
               88  NUMERIC-CHAR VALUE  0 THRU 9, '-'.  
               88  ALPHA-CHAR   VALUE 'A' THRU 'Z',
                                      'a' THRU 'z',
                                     X'80' THRU X'9A'.
               88  CR           VALUE X'0D'.
               88  LF           VALUE X'0A'.

The file is attached - contains Hebrew text.

Thanks for any help.

[Vincent (Bryan) Coen]

Did try and read it but its a code page type that I do not have in my system so cannot pass it to my wife who is reasonably fluent but really this is not important. Just as well as I use Linux.

So back to your problem : - before read each record flood the 01 record with say with high-values (you might need to change this value to a ALL X"FF" or similar).

Then read each RPN-CHAR (A : 1) testing for high-value and if so break out of perform i.e.,

::cobolfree
READ-Record.

     move high-values to  PRN-LiNE.
     read rpm-in
           at end go to we-are-done-or-whatever.

     perform varying A from 1 by 1 until A > 144   { change size of 
field two more cars ie 146).
                                   or  PRN-CHAR (A) = high-value

          *>    now do what ever processing you need on the character
     end-perform


we-are-done-or-whatever.
     close rPN-IN output-file
     goback.

As the file contains LS  records the EOL is immaterial as pre-loading high-value before each read you only need to check for that. You just have to increase the record size a few bytes longer.

Warning if you are using any of the Ivrit variants that are double byte then you have to double the size of the record but these are fairly obscure ie not easy to find anyway.

Easy.

[Anonymous]

A DOS printer file contains printer control characters other than CR and LF.
Control characters such as new page (NP x'0C'), tabs (HT x'09') and SUB x'1A'.
So ORGANIZATION IS LINE SEQUENTIAL is cumbersome and not really suitable to parse the sample file.

Perhaps it would be better to read the whole file in to storage and then parse the data.
This would make scraping the required data simpler.

In the sample file is defined as follows:

• each new page begins with CR,NP
• max line length is 150 + 2 (CR,LF)
• each page contains 7 header lines delimited by CR,LF
• no page footer
• some lines contain tabs
• detail lines, except last, are double spaced delimited by CR,LF,CR,LF
• report end is delimited by CR,x'1A'
  
  

Some sample code:

::cobolfree
SELECT PRN-IN ASSIGN TO file-in-prn ORGANIZATION IS SEQUENTIAL.

FD  PRN-IN
01 PRN-IN-data X(65536).

01 data-buffer.
  data-buffer-char X occurs 65536 times.

move X"FF" to data-buffer.
open PRN-IN.
read PRN-IN into data-buffer.
close PRN-IN. 

move 1 to index1
move 2 to index2
perform a100-parse-data until end-of-data or index1 > 65535

a100-parse-data.
*> check for end of report
if data-buffer-char(index1) equals CR and data-buffer-char(index2) equals x'1A'
then
  move true to end-of-data
else
*> check new page
  ...
  if index2 > 65536
  then
    move true to end-of-data
    move true to data-overflow
  end-if
end-if

Hope this helps.