Re: The case of outnumbered numerator

In the ultimate end, Paul.Roy is correct. Nothing in the OP suggests that
he cannot redesign the table, it reads to me more that he's looking for a
trick to avoid having to do so, and having to recompile everything that
touches the affected tables. Of course it depends on specifics, but the
effort to play hex tricks with the number may take more time to code, test
(and TEST AND TEST) and implement than to just do it the right way.

I am living a similar nightmare right now dealing with Y2K40 in this shop I
started working in a few months ago. They kicked the can down the road
back in the 90's, when they did a half-hearted (yet still significant)
effort on Y2K, and now they're going to pay again, because they failed to
deal with the 1940-2039 window IBM gave us with 2-digit years.

- Dan

On Wed, Jun 17, 2015 at 5:05 PM, <paul.roy@xxxxxxx> wrote:

My 5 cents...
In my career, I have had to solve some problem like this by tricks as
suggested (using base 16 or 36, mixing letters and numbers, using
alternate collating sequence, splitting data bases, etc... )
This is typically a design problem, and IMHO it can only be addressed by a
design change.. not by a programming trick.

Actually I just believe that this is the wrong way.
In 2015, shouldn't a developer just suggest : ALTER TABLE ... ALTER
COLUMN ... SET DATATYPE CHAR(7)
If the application cannot afford that sort of change in the data base...
the application needs "modernization" or will probably not survive the
next design problem..

My conviction/fear is that if you implement one of the proposed trick
(base 16/36,etc..) the application will die and be totally replaced within
maximum 3-5 years...
Making an application flexible/easy to adapt quickly is what I call
"modernization"
So my recommendation is : Modernize your application before your counter
reaches 999999 and do not invest in programming tricks unless you agree to
decommission your application in 3 years.

Paul




From: Jon Paris <jon.paris@xxxxxxxxxxxxxx>
To: Midrange-L Midrange-l <midrange-l@xxxxxxxxxxxx>
Date: 17/06/2015 21:57
Subject: Re: The case of outnumbered numerator
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>



I was suggesting that perhaps going all the way to base 36 was overkill
John and that base 16 might provide a satisfactory, and far easier to
implement, alternative.

My understanding of OPs original request was that he knew 99999? wasn?t
big enough and was wondering about 0-Z as an alternative. I don?t think he
said it was a requirement.


Jon Paris

www.partner400.com
www.SystemiDeveloper.com

On Jun 17, 2015, at 1:17 PM, John Yeung <gallium.arsenide@xxxxxxxxx>
wrote:

On Wed, Jun 17, 2015 at 12:46 PM, Jon Paris <jon.paris@xxxxxxxxxxxxxx>

wrote:

Another option that could be simply implemented using IBM APIs is to us

a hex key. Maintain internally as integer, add 1 and convert to hex for
output.

You expand the space from base 10 to base 16, but OP is looking to
expand to base 36. And to maintain EBCDIC sort order, which your
suggestion doesn't do either.

John Y.
--

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.