RE: Stupid question

["Stone, Joel" <Joel.Stone@xxxxxxxxxx>]

Couldn't SQL handle this task simpler and potentially process faster?

Something like Delete from 2nd file where (join =)

RTVMBRD to get the record length &R and CRTPF reclen(&R) then ovrdbf & run SQL

?



-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Pat Barber
Sent: Tuesday, May 01, 2012 2:48 PM
To: Midrange Systems Technical Discussion
Subject: Re: Stupid question

I would do as Paul mentioned with internal files and single field for 
each record.

I would then write a very simple matching records program with a
M1 indicator on each record.

Very old but very effective way to do what you are trying to do.

It would very fast compared to all that chaining.

Maybe  10-15 lines of code.

On 5/1/2012 9:16 AM, John McKee wrote:
> I posted this (I think - never get emails from posting) to the RPG
> list.  Never received a reply.
>
> I have two files.  Different names, formats, and field names.
>
> One is produced early in the day.
>
> I need to remove records from the second file that are identical to
> those in the first file.  Thus, second file only has new and changed
> records.
>
> I am stuck on how to set up a comparison of the entire record from
> each file.  I know I have done this, many years ago, but my brain is
> just too foggy.  I just don't want to have an IF with 90+ field
> comparisons.
>
> Logic is simple - read file 2, attempt to chain to file 1.  No chain -
> a new record.  If chain is successful, compare the two records.  If
> same, delete record from file 2.
>
> I can't directly compare the record formats.  How do I define a single
> field over each record?
>
> I tried I-Specs - File name followed by single field name..
>
> Dumb question.  Thanks for a nudge.
>
> John McKee