RE: accessing the records in a history file -- MIDRANGE-L

Title: RE: accessing the records in a history file

I see your point

However; it has worked well here, the history files have been basically stable for the last five years, besides, anything added to the files beginning today, we know that we do not have the data to put there in the old files - about two years ago we did add a few new fields - to the end of the physical file, the old files where left with level check *no, and they have been excluded from the newer reports since the data was never there in the first place. If you have the data to place in the history file, why would you need it there anyhow has been my philosophy - I hate to clutter up dasd with duplicate data "if" I can avoid it.

We did try to merge these files, the logicals and high usage of the files killed our little box, the detail has 9 logicals already, and the whole challenge has been for this new staff to live with, since we do not have the time nor can we make a ROI justification for doing things right after the previous team of professionals left in mass - most of them came from a S/34-S/36 background; and it shows in all the home grown code and even how the AS/400 was set up - ROI just is not there, so we clean up and make corrections as we can - some will never happen, some will have to even if management does not want it to...

As with all suggestions, comments, and experience contained with-in this list and our minds -

you must first see if it applies to where you are now,

the corporate environment

the desired results

the time frame you have been given

the funds you are allotted

the tools you have

the hardware you have to live with (wouldn't it be nice to be on the max config of an AS/400)

the people you must deal with

and lastly your comfort level to do the right thing for your company....

just another two cents, ahhh possibly a dime ;)

-----Original Message-----
From: owner-midrange-l@midrange.com [mailto:owner-midrange-l@midrange.com]On Behalf Of Leland, David
Sent: Thursday, December 21, 2000 1:10 PM
To: 'MIDRANGE-L@midrange.com'
Subject: RE: accessing the records in a history file

Hmmm, are you sure this is such a good idea? We've been there and we decided to merge all the yearly files into one and create indexes (logicals and Encoded Vector Indexes) over certain key fields (date, customer #, order number, etc.). With the multiple files, whenever a change to the file was made, we had to make sure we made it for all the files. Also, it was more difficult creating reports which accessed multiple year's data.

Dave

-----Original Message-----
From: Mark A. Manske [mailto:mmanske@minter-weisman.com]
Sent: Thursday, December 21, 2000 12:48 PM
To: MIDRANGE-L@midrange.com
Subject: RE: accessing the records in a history file

Don't tell management this(becaue it will appear you have removed records)
Split up the file, create LFs over date,
change pgms/qrys accessing file to go to proper one(s)
depending on selection criteria.

We have similar issue here, sales history detail files
by year, each year with from 13 million to 15 million records
and we are still on an "olde" 620 with lots of dasd...

Development may take longer due to accessing multiple files,
but run time will drop drastically "if" they are looking at
say only one year.

Just my two cents....

HTH

Mark A. Manske
[mailto:mmanske@minter-weisman.com]
Sr. Project Lead
Minter-Weisman

-----Original Message-----
From: owner-midrange-l@midrange.com
[mailto:owner-midrange-l@midrange.com]On Behalf Of booth@martinvt.com
Sent: Thursday, December 21, 2000 10:44 AM
To: midrange-l@midrange.com
Subject: accessing the records in a history file

I'd like some insight from folks on this scenario: There is a history
file of records dating from 1994. Records are added each day and now
there are 3.6 million records in the file. The odds of anyone going back
to look up something in 1994 are slim but management won't purge any
records. Obviously the newest records are at the end of the file. All of
the logicals have the customer ID number as the first key. One logical has
a key of the customer ID and then date and time.

Each business day there is at least one and sometimes five or more reports
written for all activity during some recent time period (a workshift, a
day, a week, whatever).

This process is becoming a performance pig; additionally there is a need
for more function. What sorts of suggestions do you people see as
possible enhancements to this? SQLRPGLE? OPNQRYF? a new Logical on date
and time alone? #GSORT?

Any suggestions are appreciated. thanks in advance.

_______________________
Booth Martin
Booth@MartinVT.com
http://www.MartinVT.com
_______________________
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator:
david@midrange.com
+---

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---