IBM Mainframe Forum

by **IBM MINAKSHI** » Sat Jan 05, 2013 12:53 am

Hi,
Please help me in getting answer for a question related to file processing in a COBOL program and create an output file. Input file has 10 message fields, i need to write an output records for 5 message. i.e. if file has data for only message1 field then need to write one o/p records, if file has data for message1,message2 then need to write two o/p records and so on. means everytime i need to check whether multiple error messages present or not. in detail-
if Record has multiple Error messages then we have to check and process only 1st 5 error messages
I. For first error message, i need to write source, Account No, id,ccode,message1 to the o/p file.
II. For subsequent 5 error messages, need to write only the error message on the next line leaving the other fields (source, account nu, ID, ccode) blank.
my input file layout is -

source x(3)
Account number length x(15)
id length x(8)
ccode x(3)
message1 x(30)
message2 x(30)
message3 x(30)
message4 x(30)
message5 x(30)
message6 x(30)
message7 x(30)
message8 x(30)
message9 x(30)
message10 x(30)

by **Akatsukami** » Sat Jan 05, 2013 12:57 am

Are we to presume that if, e.g., MESSAGE2 is non-blank, MESSAGE1 must also be non-blank?

by **IBM MINAKSHI** » Sat Jan 05, 2013 1:11 am

Akatsukami wrote:Are we to presume that if, e.g., MESSAGE2 is non-blank, MESSAGE1 must also be non-blank?

Hi,
Message1 will always have a value.. but message2....10 may or may not have a value. becoz of that every time we have to check whether message 2..5 has data or not. if yes then we have to write corresponding records..

by **dick scherrer** » Sat Jan 05, 2013 1:41 am

Hello and welcome to the forum,

Suggest you post a few sets of sample input data and the output you want what those sample sets are processed.

You have answered the literal question about message1 and message2. I beileve what is needed is the answer to if any messageN has a value, must all of the messages lower than messageN have a value or may there be unfilled messages throughout?

The sample sets should show this.

by **Akatsukami** » Sat Jan 05, 2013 2:50 am

As it happens, Minakshi-chan, I am not a COBOL programmer, at least not in recent decades. (I know, you were hoping that I'd write the program for you. Life is cruelly disappointing at times, my child.)

However, I did write this fine PL/I program:

Select all

m45: proc options (main) reorder;

%dcl (true, false, pagesize) char;
%true = '''1''B';
%false = '''0''B';
%pagesize = 34;

dcl tulin file record input,
tulout file record output;

dcl 1 input_record unal,
2 source char (3),
2 account_number char (15),
2 id char (8),
2 ccode char (3),
2 message (10) char (30);

dcl 1 output_header_1 unal,
2 * char (39) init(' '),
2 * char (31) init('KatsuSoft'),
2 * char (5) init('Page'),
2 oh_page# pic '(4)z9';

dcl 1 output_header_2 unal,
2 * char (5) init('Src'),
2 * char (17) init('Account Number'),
2 * char (10) init('ID'),
2 * char (5) init('Ccd'),
2 * char (43) init('Message');

dcl 1 output_header_3 unal,
2 * char (5) init((3)'='),
2 * char (17) init((15)'='),
2 * char (10) init((8)'='),
2 * char (5) init((3)'='),
2 * char (43) init((30)'=');

dcl 1 output_detail unal,
2 source char (5),
2 account_number char (17),
2 id char (10),
2 ccode char (5),
2 message char (43);

dcl eof_tulin bit init(false),
blank_line char (80) init(' '),
i fixed bin (31),
page# fixed bin (31) init(1),
lineage fixed bin (31) init(99);

on endfile (tulin) eof_tulin = true;

read file (tulin) into (input_record);

do while (¬eof_tulin);
output_detail.source = input_record.source;
output_detail.account_number = input_record.account_number;
output_detail.id = input_record.id;
output_detail.ccode = input_record.ccode;
output_detail.message = input_record.message (1);
call do_output;
output_detail.source = ' ';
output_detail.account_number = ' ';
output_detail.id = ' ';
output_detail.ccode = ' ';

do i = 2 to 5 while (input_record.message(i)¬=' ');
output_detail.message = input_record.message (i);
call do_output;
end;

read file (tulin) into (input_record);
end;

do_output: procedure;

if (lineage>pagesize) then do;
oh_page# = page#;
page# += 1;
write file (tulout) from (output_header_1);
write file (tulout) from (output_header_2);
write file (tulout) from (output_header_3);
write file (tulout) from (blank_line);
lineage = 4;
end;

write file (tulout) from (output_detail);
lineage += 1;
end do_output;
end m45;

which, given the input

Select all

CBL012345678901234ABCD 123This is a test
PLI123456789012345defg 456This is another test Katie is a pest You're welcome, Minakshi

produces the output

Select all

KatsuSoft Page 1
Src Account Number ID Ccd Message
=== =============== ======== === ==============================

CBL 012345678901234 ABCD 123 This is a test
PLI 123456789012345 defg 456 This is another test
Katie is a pest
You're welcome, Minakshi

You may take this as pseudo-code, or we can discuss the COBOL statements corresponding to the PL/I, your choice.

by **IBM MINAKSHI** » Sat Jan 05, 2013 9:08 am

Thanks Akatsukami...

this is what i want to give as input and write as output but thru a Cobol Program.. how can i find Cobol Program code equal to PL/I code.. Can we discuss the COBOL Statements.

by **IBM MINAKSHI** » Sat Jan 05, 2013 9:42 am

Hi dick scherrer... Thanks

Akatsukami has already given sample input and output data... i want to give input data and expecting output data as..
Sample Input is-

Select all

CBL012345678901234id0001idAAAThis is a message1
CBL012345678909876id0002idBBBThis is a message1 This is a message2 This is a message3
PLI112345678901234id0003idCCCThis is a trail1 This is a trail2 This is a trail3 This is a trail4 This is a trail5 This is a trail6 This is a trail7 This is a trail8 This is a trail9 This is a trail10

and Sample output should be -

Select all

Test Report Page 1
Src Account Number ID Ccd Message
=== =============== ======== === ==============================

CBL 012345678901234 id0001id AAA This is a message1
CBL 012345678901234 id0002id BBB This is a message1
This is a message2
This is a message3
PLI 112345678901234 id0003id CCC This is a trail1
This is a trail2
This is a trail3
This is a trail4
This is a trail5

Thanks All..

Code'd

by **Akatsukami** » Sat Jan 05, 2013 8:52 pm

We shall discuss this, Minakshi-chan. First, though, note that you ought to enclose any information where alignment is significant (and it is very significant in both the input data and the report) in Code tags, so as to preserve that alignment. Note the differences in the appearance of my report and your mock-up.

Now, to discussion. Leaving aside for moment the variable declarations and overhead, the logic begins with:

Select all

on endfile (tulin) eof_tulin = true;

read file (tulin) into (input_record);

The second statement is of course a mere I/O statement; the file tulin is being read into the variable input_record. In COBOL, the file definition (FD) must include some sort of record definition. However, this definition represents the actual I/O buffer; as the data in the buffer are volatile, you should always read into and write from working storage, so as to be sure the data are available when needed.

In PL/I I simply write

Select all

dcl tulin file record input,

whereas in COBOL you need a more complex definition. As I said supra, it has been many years since I programmed in COBOL, but IIRC the definition would be:

Select all

ENVIRONMENT DIVISION.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
SELECT INPUT-FILE ASSIGN TO TULIN.
[...]
DATA DIVISION.
FILE SECTION.
FD INPUT-FILE
DATA RECORD IS INPUT-BUFFER.
01 INPUT-BUFFER PIC X(329).

In the FILE-CONTROL paragraph, you use the SELECT statement to associate a COBOL internal file, that I have named INPUT-FILE, with a external file, TULIN; in MVS this will be the name of the DD statement defining the actual data set in the JCL. In the FILE section, you then create a FD for the internal file.

Note that I have defined a on endfile statement for tulin. When a PL/I program is compiled, object code is generated so that when the end of the data set associated with tulin is reached, control is transferred to the on statement and then, when it executes, returns to the point of origin. In COBOL, this must be done explicitly, so that

Select all

read file (tulin) into (input_record);

Becomes

Select all

READ INPUT-FILE INTO INPUT-RECORD
AT END MOVE 'Y' TO EOF-INPUT-FILE.

where, of course, EOF-INPUT-FILE is a PIC X variable in working storage, initialized to something other than 'Y' ('N' would be conventional and easy to interpret, but not absolutely necessary).

That’s enough for one morning. Digest that, and we shall then continue.

by **IBM MINAKSHI** » Sat Jan 05, 2013 9:24 pm

Thanks Akatsukami for explaining very nicely.. Also i will always keep in mind that Report Alignment should be proper and exact..

I understood your conversion very clearly.. Thanks again..
For Do while logic.. do we need to put array replacement in COBOL?

by **NicC** » Sat Jan 05, 2013 10:11 pm

Do While is Perform Until
An ARRAY in PL/1 is a TABLE in Cobol

An assignment in PL/1 is

Select all

a = b

which in Cobol would be

Select all

MOVE B TO A

or

Select all

A = B + C

in PL/1 is

Select all

COMPUTE A = B + C

in Cobol
If you do not have a pl/1 manual available (and you do via the links on this page) use the Rexx Language reference as it is more PL/1-like and you will be using Rexx throughout your career.

IBM Mainframe Forum

COBOL File Processing Position wise

COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise

Re: COBOL File Processing Position wise