RE: **SPAM** Re: Modular advice

First I'd suggest a 2 pronged attack...

Program A does a check Record v. Record (not at field level)
  if Record <> Record then call Program B to find the mismatched data.

This doesn't eliminate the need for all the code but will definitely
speed up performance I'd think.  Maybe even catch the buffer positions
that are different & do a lookup for the field name in the system tables
(QADBIFLD) based on the file / lib / buffer positions.  It might be
faster to try that approach ( & less code...)

HTH & GL!!!


Thanks,
Tommy Holden


-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx
[mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of Tony Carolla
Sent: Tuesday, January 10, 2006 3:31 PM
To: RPG programming on the AS400 / iSeries
Subject: **SPAM** Re: Modular advice

Thanks Carel.  I should have been more specific.

There are a set of source documents that represent 300 accounts.  The
students will each key all 300 of them into the test library.  The
instructor has entered the data in these files already, and I made
copies of
them in another library ("model" files).

The students will enter the data, and my programs will check what they
entered against what is in the model files, that the instructor
previously
populated.

So there are seven files that contain data (acct master, claim master,
trans
detail, etc.).  They are comprised of around three hundred fields of
data,
each of which must be tested against the known-good "master" files that
the
instructor populated previously.

I haven't sent any specific record layouts, etc. because at this point,
I am
in the planning phase.  I am trying to envision it, and all I can see is
writing a monolothic program that simply checks each field,
sequentially,
and writes an output record containing the field name and the
correct/incorrect answers to an output file.  This would result in 317
chunks of code, one after the other, and a nightmare to update later
(even
for me ;-) )

But I think that creating 317 modules, one for each field, isn't really
any
better.  Thanks in advance for any help in structuring this beast.

On 1/10/06, Carel Teijgeler <coteijgeler@xxxxxxxxx> wrote:
>
> Tony,
>
> Unfortunately, you do not tell what the files to be checked have for
> format.


--
"Enter any 11-digit prime number to continue..."
"In Hebrew SQL, how do you use right() and left()?..." - Random Thought
"If all you have is a hammer, all your problems begin to look like
nails"