Cobol - Cobol (Coding problem)

Author

Cobol (Coding problem)

Rick

2005-01-22, 3:55 am

Hi,
I have got some coding problem and appreciate someone could help advise
on how it should be done. Quite complicated though.Tks in advance.

Scenario (1)
============
1) If there is only 1 record only with version 00,
then if date1 field = non-current date
and date2 field = current date, o/p to file(B)

In short, record pattern is :
verno date1 date2
xxx xxx xxx 00 non-current date current date => o/p to
file(B)

Scenario(2)
============
2) However If there are multiple records(same key) with
version 00, 01, 02 etc, have to check that
if it is ver00 and date(1) field = non-current date
and date(2) field = current date, o/p to file(B)

For the rest of the record with same key (ie version 01,02 etc)
if date(1) field = current date or that date(2) field is
current or non-current date, o/p to file(A)

In short, record pattern is :
verno date1 date2
xxx xxx xxx 00 non-current date current date => o/p to file(B)
xxx xxx xxx 01 current date current date => o/p to file(A)
xxx xxx xxx 02 current date non-current date => o/p to file(A)

(P/S: For o/p to file(B), only allow for version 00 records only)

How should I code for something that is able to perform the above??

I have some coding and towards the end, I coded like

If group-change => group change refer to the group of any xxx field
above
if date(2) field = current date
o/p to file(B)
end-if
end-if

** It work for scenario(1) only but got bug in scenario(2). Can advise
how the coding should be done ??

** I thought of coding to check that if version no. not = spaces, then
perform
something, but if you had ver00 or version01,02,etc, all these are consider
not
= spaces.They are suppose to perform different things for version that is
00,
and the rest of the version 01,02 etc.Not sure how I should code it.

Pete Dashwood

2005-01-22, 8:55 am

"Rick" <leng1@bigfoot.com> wrote in message
news:cssq0a$95o$1@reader01.singnet.com.sg...
> Hi,
> I have got some coding problem and appreciate someone could help advise
> on how it should be done. Quite complicated though.Tks in advance.
>
>
> Scenario (1)
> ============
> 1) If there is only 1 record only with version 00,
> then if date1 field = non-current date
> and date2 field = current date, o/p to file(B)
>
> In short, record pattern is :
> verno date1 date2
> xxx xxx xxx 00 non-current date current date => o/p to
> file(B)
>

I assume that if there is only 1 record with version 00, and date 1 field is
non-current, and date 2 field is also non-current,
then it goes to file A, right?
>
> Scenario(2)
> ============
> 2) However If there are multiple records(same key) with
> version 00, 01, 02 etc, have to check that
> if it is ver00 and date(1) field = non-current date
> and date(2) field = current date, o/p to file(B)
>
> For the rest of the record with same key (ie version 01,02 etc)
> if date(1) field = current date or that date(2) field is
> current or non-current date, o/p to file(A)
>
>
>
> In short, record pattern is :
> verno date1 date2
> xxx xxx xxx 00 non-current date current date => o/p to
file(B)
> xxx xxx xxx 01 current date current date => o/p to
file(A)
> xxx xxx xxx 02 current date non-current date => o/p to
file(A)
>
>
> (P/S: For o/p to file(B), only allow for version 00 records only)
>
>
>
> How should I code for something that is able to perform the above??
>
> I have some coding and towards the end, I coded like
>
> If group-change => group change refer to the group of any xxx field
> above
> if date(2) field = current date
> o/p to file(B)
> end-if
> end-if
>
> ** It work for scenario(1) only but got bug in scenario(2). Can advise
> how the coding should be done ??
>
> ** I thought of coding to check that if version no. not = spaces, then
> perform
> something, but if you had ver00 or version01,02,etc, all these are
consider
> not
> = spaces.They are suppose to perform different things for version that is
> 00,
> and the rest of the version 01,02 etc.Not sure how I should code it.
>
>

It is an interesting problem, Rick and you have stated it well.

There are a number of red herrings and you were wise to build truth tables.

Ignoring what is not important, I believe the following should do it (unless
there are other conditions you haven't mentioned.)

WS

01 date-1-flag pic x value zero.
88 date-1-current value '1'.

01 date-2-flag pic x value zero.
88 date-2-current value '1'.

01 today.
12 today-yyyy pic x(4).
12 today-mm pic xx.
12 today-dd pic xx.

PD
.....
* set up the today field once in initialization, unless you are likely to
run
* this process overnight and date could change at midnight.
*
* If your compiler supports functions use function (CURRENT-DATE) otherwise
* you may need to accept TODAYS-DATE... check your COBOL manual to see what
works for you.
I
move function (CURRENT-DATE) to today *> may need some fiddling to
ensure formats match

*Apply the following code to ALL records in each key group as each record is
fetched...

*You might want to detect a group change for other reasons, but you don't
need to on the basis of what you told us.

move zero to date-1-flag date-2-flag
if date-1-field = today
set date-1-current to TRUE
end-if
if date-2-field = today
set date-2-current to TRUE
end-if
if version = '00'
AND
NOT date-1-current
AND
date-2-current
output to file B
else
output to file A
end-if

etc.

HTH,

Pete.

Lueko Willms

2005-01-22, 8:55 am

.. On 22.01.05
wrote leng1@bigfoot.com (Rick)
on /COMP/LANG/COBOL
in cssq0a$95o$1@reader01.singnet.com.sg
about Cobol (Coding problem)

I allow myself to rearrange the quoted descriptions a little bit

l> Scenario(1)
l> ============
l> 1) If there is only 1 record only with version 00,
l> then
l> if date1 field = non-current date
l> and date2 field = current date,
THEN
l> o/p to file(B)

l> Scenario(2)
l> ============
l> 2) However If there are multiple records(same key) with
l> version 00, 01, 02 etc, have to check that

l> if it is ver00
l> and date(1) field = non-current date
l> and date(2) field = current date,
THEN
l> o/p to file(B)

Frankly, I do not see any difference between the two scenarios,
unless you say that the in the case of _multiple_ occurences of
Version 00 the second and subsequent records with Version ßß have to
be treated differently.

You can use the same conditional for both scenarios, as coded above
by yourself (well, some syntactic accomodations, like "Date1 NOT =
Current-Date", e.g.).

You might of course try something more fancy, and use EVALUATE:

EVALUATE
Version, ALSO Date1 = Current-Date, ALSO Date2 = Current-Date

WHEN '00', ALSO FALSE, ALSO TRUE
WRITE File-B-Record FROM INput-File-Record
OTHERWISE
WRITE File-A-Record FROM Input-File-Record
END-EVALUATE

But for a binary choice, you might as well use IF .. THEN .. ELSE..
END-IF.

HAve fun!
Lüko Willms http://www.willms-edv.de
/--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --

Die Superklugheit ist eine der verächtlichsten Arten von Unklugheit. -G.C.Lichtenberg

Rick

2005-01-22, 3:55 pm

Hi,
There is different for 2nd scenario

Record pattern is :
verno date1 date2
xxx xxx xxx 00 non-current date current date => o/p to
file(B)
xxx xxx xxx 01 current date current date => o/p
to file(A)
xxx xxx xxx 02 current date non-current date => o/p to
file(A)

For ver00, date2 must be current date, then o/p to file(B)
For all other version of the SAME group(xxx), so long date1 is current date,
o/p to file(A)
Of course, I do not know how may version there will be so may need to use
'PERFORM... UNTIL'
to write out all the records of the same group (ver01,02,03 etc) to file(A)
so long date1 field
is current date but not sure how to go about coding it using PERFORM....
UNTIL

"Lueko Willms" <l.willms@jpberlin.de> wrote in message
news:9PKJzgN9flB@jpberlin-l.willms.jpberlin.de...
>. On 22.01.05
> wrote leng1@bigfoot.com (Rick)
> on /COMP/LANG/COBOL
> in cssq0a$95o$1@reader01.singnet.com.sg
> about Cobol (Coding problem)
>
>
> I allow myself to rearrange the quoted descriptions a little bit
>
>
> l> Scenario(1)
> l> ============
> l> 1) If there is only 1 record only with version 00,
> l> then
> l> if date1 field = non-current date
> l> and date2 field = current date,
> THEN
> l> o/p to file(B)
>
>
> l> Scenario(2)
> l> ============
> l> 2) However If there are multiple records(same key) with
> l> version 00, 01, 02 etc, have to check that
>
> l> if it is ver00
> l> and date(1) field = non-current date
> l> and date(2) field = current date,
> THEN
> l> o/p to file(B)
>
>
> Frankly, I do not see any difference between the two scenarios,
> unless you say that the in the case of _multiple_ occurences of
> Version 00 the second and subsequent records with Version ßß have to
> be treated differently.
==> Yes, they are supposed to be treated differently

>
> You can use the same conditional for both scenarios, as coded above
> by yourself (well, some syntactic accomodations, like "Date1 NOT =
> Current-Date", e.g.).
>
> You might of course try something more fancy, and use EVALUATE:
>
>
>
> EVALUATE
> Version, ALSO Date1 = Current-Date, ALSO Date2 = Current-Date
>
> WHEN '00', ALSO FALSE, ALSO TRUE
> WRITE File-B-Record FROM INput-File-Record
> OTHERWISE
> WRITE File-A-Record FROM Input-File-Record
> END-EVALUATE
>
>
> But for a binary choice, you might as well use IF .. THEN .. ELSE..
> END-IF.
>
>
> HAve fun!
> Lüko Willms http://www.willms-edv.de
> /--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --
>
> Die Superklugheit ist eine der verächtlichsten Arten von
> Unklugheit. -G.C.Lichtenberg

Rick

2005-01-22, 3:55 pm

"Pete Dashwood" <dashwood@enternet.co.nz> wrote in message
news:35em0jF4in9cvU1@individual.net...
>
> "Rick" <leng1@bigfoot.com> wrote in message
> news:cssq0a$95o$1@reader01.singnet.com.sg...
>
> I assume that if there is only 1 record with version 00, and date 1 field
> is
> non-current, and date 2 field is also non-current,
> then it goes to file A, right?
===> No, basically, it will be ignored for these record. For only 1 record
with version00, date2 field must be current date, then o/p to
file(B).
This is to ensure that I pick current day record but I need to
determine base on the 2 date field to o/p to which file

> file(B)
> file(A)
> file(A)
> consider
>
> It is an interesting problem, Rick and you have stated it well.
>
> There are a number of red herrings and you were wise to build truth
> tables.
>
> Ignoring what is not important, I believe the following should do it
> (unless
> there are other conditions you haven't mentioned.)
>
>
> WS
>
> 01 date-1-flag pic x value zero.
> 88 date-1-current value '1'.
>
> 01 date-2-flag pic x value zero.
> 88 date-2-current value '1'.
>
> 01 today.
> 12 today-yyyy pic x(4).
> 12 today-mm pic xx.
> 12 today-dd pic xx.
>
> PD
> ....
> * set up the today field once in initialization, unless you are likely to
> run
> * this process overnight and date could change at midnight.
> *
> * If your compiler supports functions use function (CURRENT-DATE)
> otherwise
> * you may need to accept TODAYS-DATE... check your COBOL manual to see
> what
> works for you.
> I
> move function (CURRENT-DATE) to today *> may need some fiddling
> to
> ensure formats match
>
>
> *Apply the following code to ALL records in each key group as each record
> is
> fetched...
>
> *You might want to detect a group change for other reasons, but you don't
> need to on the basis of what you told us.
>
> move zero to date-1-flag date-2-flag
> if date-1-field = today
> set date-1-current to TRUE
> end-if
> if date-2-field = today
> set date-2-current to TRUE
> end-if
> if version = '00'
> AND
> NOT date-1-current
> AND
> date-2-current
> output to file B
> else
> output to file A
> end-if
>
> etc.
===> Hi, does the above code o/p for the rest of the version (for
ver01,02,03 etc) of the same group
to file(A).As I do not know how many version there will be,
basically so long for version00
and date2 field is current date, then o/p to file(B), all
subsequent version (01,02,03 etc)
of the same group, will need to o/p to file(A) so long date(1)
field is current date.Guess, have to use
"PERFORM... UNTIL" but not sure how to code. Can help ??

Lueko Willms

2005-01-22, 8:55 pm

.. On 23.01.05
wrote leng1@bigfoot.com (Rick)
on /COMP/LANG/COBOL
in cstuvh$afc$1@reader01.singnet.com.sg
about Re: Cobol (Coding problem)

l> There is different for 2nd scenario

according to what you explained in your original message, I do not
see the slightest difference, and what you advance here, does it
neither.

l>
l> Record pattern is :
l> verno date1 date2
l> xxx xxx xxx 00 non-current date current date =>
l> o/p to file(B)
l> xxx xxx xxx 01 current date current date
l> => o/p to file(A)
l> xxx xxx xxx 02 current date non-current date =>
l> o/p to file(A)
l>
l> For ver00, date2 must be current date, then o/p to file(B)
l> For all other version of the SAME group(xxx), so long date1 is current
l> date, o/p to file(A)

You have three conditions for the record being copied to file_B
instead of file_A:

Version = '00'
AND date1 NOT = current-date
AND date2 = current-date

all input-records which do not fulfill these conditions are being
copied to file_A.

Did I miss something? Then explain.

The only loop you have to code is the one over the valid records
of the input file; the decision for copying is done via a simple
IF ... THEN ... ELSE ... END-IF

The whole program (well, the procedural part) could go like this:

--------- schnipp -----------------------------------------
PROCEDURE DIVISION.
DECLARATIVES.

*> event handlers for unrecoverable errors
*> on input or output files go here

END-DECLARATIVES.

MAIN SECTION.
BEGIN.
OPEN INPUT input-file
IF is-not-present-file in status-input-file
THEN *> minor filestat = '5'
DISPLAY "input-file is not present"
ELSE
MOVE FUNCTION(CURRENT-DATE) TO My-Current-Date
READ input-file
IF has-reached-EOF in status-input-file
THEN *> Major filestat = '1'
DISPLAY "input-file is empty"
ELSE
OPEN OUTPUT output-file-A, output-file-B
PERFORM WITH TEST AFTER
UNTIL has-reached-EOF in status-input-file
*> we now surely do have a valid record from input-file
IF Version OF input-file-record = '00'
AND date-1 OF input-file-record NOT = My-Current-Date
AND date-2 OF input-file-record = My-Current-Date
THEN
WRITE output-file-B-record FROM input-file-record
ELSE
WRITE outpuf-file-A-record FROM input-file-record
END-IF
READ input-file
END-PERFORM
CLOSE output-file-A, output-file-B
END-IF
CLOSE input-file
END-IF

slade2

2005-01-23, 3:55 am

Hi Rick,

If I read your specs correctly, this pseudo code should work:
(Don't forget the Os and Cs)

W/S.
01 NEW-REC.
05 NEW-etc ...

01 PREV-REC.
05 PREV-etc ...

01 GRP-FIN-SW PIC X VALUE 'R'.
88 RESET-GRP-FIN-SW VALUE 'R'.
88 GRP-IS-FINISHED VALUE 'F'.

PROC DIV.
Get the today's date
PERF READ-IP-REC
PERF PROCESS-GROUPS UNTIL EOF
STOP RUN
..
PROCESS-GROUPS.
You can add some statistical stuff here.
SET RESET-GRP-FIN-SW TO TRUE
PERF PROCESS-RECS UNTIL GRP-IS-FINISHED
You can add some statistical stuff here.
..
PROCESS-RECS.
EVAL NEW-VERNO
WHEN 0
IF NEW-DATE-1 = TODAYS-DATE
AND
NEW-DATE-2 NOT = TODAYS-DATE
WRITE FILE-B-REC FROM NEW-REC
END-IF
WHEN OTHER
IF NEW-DATE-1 = TODAYS-DATE
WRITE FILE-A-REC FROM NEW-REC
END-IF
END-EVAL
PERF READ-IP-REC
You can add some statistical stuff here too.
..
READ-IP-REC.
MOVE NEW-REC TO PREV-REC
READ IP-REC INTO NEW-REC AT END
SET EOF TO TRUE
END-READ
IF NEW-GROUP NOT = PREV-GROUP
SET GRP-IS-FINISHED TO TRUE
END-IF

slade2

2005-01-23, 3:55 am

I knew I needed something else in that "AT END" clause. Add this SET
stmt:

SET GRP-IS-FINISHED TO TRUE

Jack

JerryMouse

2005-01-23, 3:55 am

Lueko Willms wrote:
> . On 23.01.05
> wrote leng1@bigfoot.com (Rick)
> on /COMP/LANG/COBOL
> in cstuvh$afc$1@reader01.singnet.com.sg
> about Re: Cobol (Coding problem)
>
>
> l> There is different for 2nd scenario
>
> according to what you explained in your original message, I do not
> see the slightest difference, and what you advance here, does it
> neither.
>
> l>
> l> Record pattern is :
> l> verno date1 date2
> l> xxx xxx xxx 00 non-current date current date
> =>
> l> o/p to file(B)
> l> xxx xxx xxx 01 current date current date
> l> => o/p to file(A)
> l> xxx xxx xxx 02 current date non-current date
> =>
> l> o/p to file(A)
> l>
> l> For ver00, date2 must be current date, then o/p to file(B)
> l> For all other version of the SAME group(xxx), so long date1 is
> current
> l> date, o/p to file(A)
>
> You have three conditions for the record being copied to file_B
> instead of file_A:
>
> Version = '00'
> AND date1 NOT = current-date
> AND date2 = current-date
>
> all input-records which do not fulfill these conditions are being
> copied to file_A.
>
> Did I miss something? Then explain.
>
> The only loop you have to code is the one over the valid records
> of the input file; the decision for copying is done via a simple
> IF ... THEN ... ELSE ... END-IF
>
>
> The whole program (well, the procedural part) could go like this:
>
>
>
> --------- schnipp -----------------------------------------
> PROCEDURE DIVISION.
> DECLARATIVES.
>
> *> event handlers for unrecoverable errors
> *> on input or output files go here
>
> END-DECLARATIVES.
>
> MAIN SECTION.
> BEGIN.
> OPEN INPUT input-file
> IF is-not-present-file in status-input-file
> THEN *> minor filestat = '5'
> DISPLAY "input-file is not present"
> ELSE
> MOVE FUNCTION(CURRENT-DATE) TO My-Current-Date
> READ input-file
> IF has-reached-EOF in status-input-file
> THEN *> Major filestat = '1'
> DISPLAY "input-file is empty"
> ELSE
> OPEN OUTPUT output-file-A, output-file-B
> PERFORM WITH TEST AFTER
> UNTIL has-reached-EOF in status-input-file
> *> we now surely do have a valid record from input-file
> IF Version OF input-file-record = '00'
> AND date-1 OF input-file-record NOT = My-Current-Date
> AND date-2 OF input-file-record = My-Current-Date
> THEN
> WRITE output-file-B-record FROM input-file-record
> ELSE
> WRITE outpuf-file-A-record FROM input-file-record
> END-IF
> READ input-file
> END-PERFORM
> CLOSE output-file-A, output-file-B
> END-IF
> CLOSE input-file
> END-IF
> .
> END-IT.
> STOP RUN

The above code will inspire a beginner to take up hog farming.

How about :

Open INPUT Input-file.
READ Input-File.
PERFORM until Input-FS not = '00'
IF File-1-Input-Date ....
....
END-IF
IF File-1-Input-Date...
...
END-IF
If File-1-Input-Date...
...
END-IF
Read Input-file
END-PERFORM

CLOSE Input-File OutputA-File OutputB-File.

No compound IFs or qualified data names.

Rick

2005-01-23, 3:55 am

Tks Lueko Willms for your advise :-) I will go try it out and hopefully it
will help resolve my issue :-)
Will updates through this newsgroup again.
As I also have to cater other conditions beside this in the pgm, hope
everything go well.
As the records does not have indicator field to determine where the record
should o/p to,
it's based on "record pattern" in the input file to determine, hence not so
straight forward at times :-(

<l.willms@jpberlin.de> wrote in message
news:9PKO$z7eflB@jpberlin-l.willms.jpberlin.de...
>. On 23.01.05
> wrote leng1@bigfoot.com (Rick)
> on /COMP/LANG/COBOL
> in cstuvh$afc$1@reader01.singnet.com.sg
> about Re: Cobol (Coding problem)
>
>
> l> There is different for 2nd scenario
>
> according to what you explained in your original message, I do not
> see the slightest difference, and what you advance here, does it
> neither.
>
> l>
> l> Record pattern is :
> l> verno date1 date2
> l> xxx xxx xxx 00 non-current date current date =>
> l> o/p to file(B)
> l> xxx xxx xxx 01 current date current date
> l> => o/p to file(A)
> l> xxx xxx xxx 02 current date non-current date =>
> l> o/p to file(A)
> l>
> l> For ver00, date2 must be current date, then o/p to file(B)
> l> For all other version of the SAME group(xxx), so long date1 is current
> l> date, o/p to file(A)
>
> You have three conditions for the record being copied to file_B
> instead of file_A:
>
> Version = '00'
> AND date1 NOT = current-date
> AND date2 = current-date
>
> all input-records which do not fulfill these conditions are being
> copied to file_A.
>
> Did I miss something? Then explain.
>
> The only loop you have to code is the one over the valid records
> of the input file; the decision for copying is done via a simple
> IF ... THEN ... ELSE ... END-IF
>
>
> The whole program (well, the procedural part) could go like this:
>
>
>
> --------- schnipp -----------------------------------------
> PROCEDURE DIVISION.
> DECLARATIVES.
>
> *> event handlers for unrecoverable errors
> *> on input or output files go here
>
> END-DECLARATIVES.
>
> MAIN SECTION.
> BEGIN.
> OPEN INPUT input-file
> IF is-not-present-file in status-input-file
> THEN *> minor filestat = '5'
> DISPLAY "input-file is not present"
> ELSE
> MOVE FUNCTION(CURRENT-DATE) TO My-Current-Date
> READ input-file
> IF has-reached-EOF in status-input-file
> THEN *> Major filestat = '1'
> DISPLAY "input-file is empty"
> ELSE
> OPEN OUTPUT output-file-A, output-file-B
> PERFORM WITH TEST AFTER
> UNTIL has-reached-EOF in status-input-file
> *> we now surely do have a valid record from input-file
> IF Version OF input-file-record = '00'
> AND date-1 OF input-file-record NOT = My-Current-Date
> AND date-2 OF input-file-record = My-Current-Date
> THEN
> WRITE output-file-B-record FROM input-file-record
> ELSE
> WRITE outpuf-file-A-record FROM input-file-record
> END-IF
> READ input-file
> END-PERFORM
> CLOSE output-file-A, output-file-B
> END-IF
> CLOSE input-file
> END-IF
> .
> END-IT.
> STOP RUN
> .
> ------------------ schnapp --------------------------------
>
> You would need the three preceding DIVISIONs with all the necessary
> declarations, of course.
>
> Errors with OPEN or CLOSE or other unrecoverable errors are dealt
> with in the event-handlers in the DECLARATIVES.
>
>
> Where is the problem?
>
> I could not discern any other conditions from your explanations.
>
>
> Yours,
> Lüko Willms http://www.willms-edv.de
> /--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --
>
> Er liebte hauptsächlich die Wörter, die nicht in Wörterbüchern vorzukommen
> pflegen. -G.C.Lichtenberg

Lueko Willms

2005-01-23, 8:55 am

.. On 23.01.05
wrote leng1@bigfoot.com (Rick)
on /COMP/LANG/COBOL
in csvc3f$et5$1@reader01.singnet.com.sg
about Re: Cobol (Coding problem)

l> As I also have to cater other conditions beside this in the pgm, hope
l> everything go well.

Sure, I could only respond to the questions you asked. Good luck
with the rest. How long have you been working in COBOL, and do you
other tasks besides programming in COBOL?

l> As the records does not have indicator field to determine where the
l> record should o/p to, it's based on "record pattern" in the input
l> file to determine, hence not so straight forward at times :-(

Well, you have described a clear cut, quite simple condition based
on three data items, which can comfortably be described in one
conditional expression.

'Pattern matching' suggests something else; this is meant normally
to search for patterns in one single string of text, not fixed fields
which can be identified easily by their names.

Yours,
Lüko Willms http://www.willms-edv.de
/--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --

Er liebte hauptsächlich die Wörter, die nicht in Wörterbüchern vorzukommen pflegen. -G.C.Lichtenberg

Lueko Willms

2005-01-23, 8:55 am

.. On 22.01.05
wrote nospam@bisusa.com (JerryMouse)
on /COMP/LANG/COBOL
in 10v63v947i6g071@news.supernews.com
about Re: Cobol (Coding problem)

n> The above code will inspire a beginner to take up hog farming.

which can be a rewarding job, and which requires the ability to
attend a lot of detail, the overall look at the herd, the detailed
look at the individual animals, taking care of input and output, and
this all at the same time.

Writing a program to process a sequence of data items is child play
compared to that.

n>
n> How about :
n>
n> Open INPUT Input-file.

where are the two output files?

n> READ Input-File.
n> PERFORM until Input-FS not = '00'

a minor status not equal to zero can accompany the major status of
zero, indicating SUCCESSFUL COMPLETION. Therefore I only use the major
status. The data names for the various conditions do come, of course,
from a COPY element.

n> IF File-1-Input-Date ....
n> ....
n> END-IF
n> IF File-1-Input-Date...
n> ...
n> END-IF
n> If File-1-Input-Date...
n> ...
n> END-IF

Why three consecutive conditional executions? Is that just a lapse
of cut-and-paste, or a generalisation, or deliberate?
There is only one choice (as far as Rick has told us to solve his
problem).

n> Read Input-file
n> END-PERFORM
n>
n> CLOSE Input-File OutputA-File OutputB-File.

n> No compound IFs or qualified data names.

I have just filled my template for sequential processing with the
details of Rick's question.

The rest is the same for every processing of sequentially accessed
data items of unknown quantity.

It allows the input file being OPTIONAL, and then not present:
(which is not necessary in all cases, but for my simplicity,
I maintain only one template; also non-presence of the file
is a frequent error, which is reported in a 'soft' way
by the program itself instead of a crash by the OS);
it allows the input file to be empty, and tells the user about it,
making a difference between some processing error
and an empty file;
it cares about output files only when there is input
to be processed, avoiding useless work
it is simple :-)

Also I assume the same data description for the input file and the
two output files which it is split into, so the data items have to be
qualified, to make it simple, and the code to be understood without
any research in other parts of the program or external documentation.

Do we want to start another outgrowth of a thread on a concrete
question to a generalist debate about programming principles?

Yours,
Lüko Willms http://www.willms-edv.de
/--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --

Es regnete so stark, daß alle Schweine rein und alle Menschen dreckig
wurden. -G.C.Lichtenberg

JerryMouse

2005-01-24, 3:55 am

Lueko Willms wrote:
> . On 22.01.05
> wrote nospam@bisusa.com (JerryMouse)
> on /COMP/LANG/COBOL
> in 10v63v947i6g071@news.supernews.com
> about Re: Cobol (Coding problem)
>
>
> n> The above code will inspire a beginner to take up hog farming.
>
> which can be a rewarding job, and which requires the ability to
> attend a lot of detail, the overall look at the herd, the detailed
> look at the individual animals, taking care of input and output, and
> this all at the same time.
>
> Writing a program to process a sequence of data items is child play
> compared to that.
>
>
> n>
> n> How about :
> n>
> n> Open INPUT Input-file.
>
> where are the two output files?

Output is left to the user.

>
> n> READ Input-File.
> n> PERFORM until Input-FS not = '00'
>
> a minor status not equal to zero can accompany the major status of
> zero, indicating SUCCESSFUL COMPLETION. Therefore I only use the major
> status. The data names for the various conditions do come, of course,
> from a COPY element.

Yes it can. I, however, in over thirty years have never seen it Neither will
a beginner with a single input file that's simply trying to split the file.
There is no reason to comple a beginner to fuss with a COPY statement.

>
> n> IF File-1-Input-Date ....
> n> ....
> n> END-IF
> n> IF File-1-Input-Date...
> n> ...
> n> END-IF
> n> If File-1-Input-Date...
> n> ...
> n> END-IF
>
> Why three consecutive conditional executions? Is that just a lapse
> of cut-and-paste, or a generalisation, or deliberate?

It is deliberate in the sense that it avoids a compound IF.

> There is only one choice (as far as Rick has told us to solve his
> problem).
>
> n> Read Input-file
> n> END-PERFORM
> n>
> n> CLOSE Input-File OutputA-File OutputB-File.
>
>
> n> No compound IFs or qualified data names.
>
>
> I have just filled my template for sequential processing with the
> details of Rick's question.

Rick doesn't have a template. Rick doesn't want a template. Rick may not
even know what a template is. Nor do I. Exactly.

>
> The rest is the same for every processing of sequentially accessed
> data items of unknown quantity.
>
> It allows the input file being OPTIONAL, and then not present:
> (which is not necessary in all cases, but for my simplicity,
> I maintain only one template; also non-presence of the file
> is a frequent error, which is reported in a 'soft' way
> by the program itself instead of a crash by the OS);
> it allows the input file to be empty, and tells the user about it,
> making a difference between some processing error
> and an empty file;
> it cares about output files only when there is input
> to be processed, avoiding useless work
> it is simple :-)

It is not simple. It is convoluted, complex, and over-kill. It is the worst
of all worlds for a beginner. He's GOT the file; it can never be missing.
Likewise, this one-off program does NOT have an empty file. There's two
unnecessary excursions right there.

>
>
> Also I assume the same data description for the input file and the
> two output files which it is split into, so the data items have to be
> qualified, to make it simple, and the code to be understood without
> any research in other parts of the program or external documentation.

Where'd you get the idea that qualified data names are simple? In this
example, since the ouput files never reference their data fields, why should
the input file's data names be qualified other than to feed the construct
forced by using a standardized "template?"
>
> Do we want to start another outgrowth of a thread on a concrete
> question to a generalist debate about programming principles?

Uh, isn't that what you did?

Peter Lacey

2005-01-24, 3:55 pm

JerryMouse wrote:
>
>
> It is not simple. It is convoluted, complex, and over-kill. It is the worst
> of all worlds for a beginner. He's GOT the file; it can never be missing.
> Likewise, this one-off program does NOT have an empty file. There's two
> unnecessary excursions right there.
>

Correction: "There's" is short for "There is". So you're saying "There
is two unnecessary ...".

This causes me extreme anguish!

PL

Chuck Stevens

2005-01-24, 3:55 pm

Well, since we seem to be into criticisms of the English grammar of others,
I would have to make the argument that "There's" points back to the previous
text (singular). He could equally reasonably have written "That's" (which
would probably have been my preference). How many places can one (verbal)
index finger point at a given instant? ;-)

As my sainted mother used to say: Never try to teach your grandmother how
to milk ducks ...

-Chuck Stevens

"Peter Lacey" <lacey@mb.sympatico.ca> wrote in message
news:41F5217B.39272D4D@mb.sympatico.ca...
> JerryMouse wrote:
worst[color=darkred]
missing.[color=darkred]
>
>
> Correction: "There's" is short for "There is". So you're saying "There
> is two unnecessary ...".
>
> This causes me extreme anguish!
>
> PL

Richard

2005-01-24, 3:55 pm

> He's GOT the file; it can never be missing.

What if he has mis-specified the file name ? The file might be _there_
but he looks for it _here_.

> Likewise, this one-off program does NOT have an empty file.

Idioms are designed to cater for boundary conditions, such as empty
files. It is important to teach beginners idioms which work rather
than dismissing problems which may occur.

Howard Brazee

2005-01-24, 3:55 pm

On 24-Jan-2005, "Chuck Stevens" <charles.stevens@unisys.com> wrote:

> Well, since we seem to be into criticisms of the English grammar of others,
> I would have to make the argument that "There's" points back to the previous
> text (singular). He could equally reasonably have written "That's" (which
> would probably have been my preference). How many places can one (verbal)
> index finger point at a given instant? ;-)

I disagree. It still means "There is two unnecessary excursions right there".

Lueko Willms

2005-01-24, 3:55 pm

.. On 23.01.05
wrote nospam@bisusa.com (JerryMouse)
on /COMP/LANG/COBOL
in 10v8u702s8mlcde@news.supernews.com
about Re: Cobol (Coding problem)

on my template for processing sequences of objects serially:

n> It is not simple. It is convoluted, complex, and over-kill.

it is appropriate to the task, and really simple. One can overlook
it easily since it will not strech beyond a page, neither printed nor
displayed on screen (OK, not the Epson PX-8 I once had, with only 8
lines).

n> It is the worst of all worlds for a beginner.

Rick X. did not state that he is a beginner. How do you know?

And even if he is, shouldn't a beginner begin with correct
programs?

I think he just got entangled in a specification which led him
astray, not seeing the wood for the trees (as far as I can see based
on the way he presented his problem).

n> He's GOT the file; it can never be missing.

But maybe the program doesn't find it? Or the next cycle of
execution fails to produce the output which should be the input-file
for this step?

n> Likewise, this one-off program does NOT have an empty file.

How do you know that this is a "one-off program"?

n> There's two unnecessary excursions right there.

I am talking about safeguards against potential hazards.
[color=darkred]

n> Where'd you get the idea that qualified data names are simple?

In my experience. The qualification tells the reader right away
what the program is talking about and dealing with: With items in the
input-file-record.

n> In this example, since the ouput files never reference their data
n> fields,

How do you know that?

n> why should the input file's data names be qualified other than to
n> feed the construct forced by using a standardized "template?"

The qualification does not flow from the template, but from the
ease of documentation, and avoiding separate and therefore potentially
conflicting descriptions for the three files.

If your assumption is correct, and the compiler allows it, the
output file(s) could be declared with a record of ANY size, or
something like that. But I don't know, and I do not want to act too
much on assumptions which may be wrong.
[color=darkred]

n> Uh, isn't that what you did?

Only if there are takes. :-))

Besides, I think that would be more appropriate for this
newsgroup, than the talk about political issues. In the past w

s, I
sometimes asked myself if COBOL is still on-topic here.

Yours,
Lüko Willms http://www.willms-edv.de
/--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --

Gerade das Gegenteil tun heißt auch nachahmen, es heißt nämlich das
Gegenteil nachahmen. -G.C.Lichtenberg

Chuck Stevens

2005-01-24, 8:55 pm

I think it could also be a legitimate transformation of
[Up] there is [an illustration of] two unnecessary excursions [, ] right
there.

I would personally have preferred
That is [an illustration of] two unnecessary excursions [, ] right
there.

but I think Chomsky would have recognized the transforms that could be used
to lead to the first.

-Chuck Stevens

"Howard Brazee" <howard@brazee.net> wrote in message
news:ct3f0n$b4k$1@peabody.colorado.edu...
>
> On 24-Jan-2005, "Chuck Stevens" <charles.stevens@unisys.com> wrote:
>
others,[color=darkred]
previous[color=darkred]
(which[color=darkred]
(verbal)[color=darkred]
>
> I disagree. It still means "There is two unnecessary excursions right
there".

Lueko Willms

2005-01-25, 3:55 pm

.. On 22.01.05
wrote L.Willms@JPBERLIN.de (Lueko Willms)
on /COMP/LANG/COBOL
in 9PKO$z7eflB@jpberlin-l.willms.jpberlin.de
about Re: Cobol (Coding problem)

A caveat to my proposal:

LW> MOVE FUNCTION(CURRENT-DATE) TO My-Current-Date

this can be a problem in this decision:

LW> IF Version OF input-file-record = '00'
LW> AND date-1 OF input-file-record NOT = My-Current-Date
LW> AND date-2 OF input-file-record = My-Current-Date

since FUNCTION(CURRENT-DATE) does return not only the date, but
also the time of day in one structure, and the time of day will quite
certainly not match the value given in the input-file. It does
probably only carry a calendar date.

So either My-Current-Date has to be a 8-character field, truncating
the result of FUNCTION(CURRENT-DATE) to the date only, or one has to
use ACCEPT My-Current-Date FROM DATE or do possibly shuffle the year,
month, and day parts, so that it corresponds to the form the date is
represented in the file.

Sorry about any confusion my proposal may have caused in this
point.

Yours,
Lüko Willms http://www.willms-edv.de
/--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --

Nach einem dreißigjährigen Krieg mit sich selbst kam es endlich zu einem Vergleich, aber die Zeit war verloren. -G.C.Lichtenberg

Sponsored Links