RE: Allow *ZERO in a date (and timestamp) field? IBM?

For a "from-scratch" application, I agree this is how it should be done. One
problem is most RPG programmers have not encountered nullable fields before.
Can/does RPG/III support null capable fields? I know RPG/IV has %nulind to
check against. Convert all programs to RPG/IV, recompile with
ALWNULL(*USRCTL), and bracket date operations with %nulind. I agree
technically that one shouldn't use a "real" date to represent no date, but
with as much legacy code with 6,0 or 8,0 numeric with 0 for no date, I think
the "0001-01-01" is reasonable.

Loyd

-----Original Message-----
From: R. Bruce Hoffman, Jr. [mailto:rbruceh@attglobal.net]
Sent: Thursday, August 22, 2002 9:00 AM
To: midrange-l@midrange.com
Subject: Re: Allow *ZERO in a date (and timestamp) field? IBM?


----- Original Message -----
From: "William A.(Tony) Corbett" <corbett@ASRESOURCES.COM>
To: "MIDRANGE-L@midrange. com" <MIDRANGE-L@midrange.com>
Sent: Thursday, August 22, 2002 9:05 AM
Subject: Allow *ZERO in a date (and timestamp) field? IBM?


> If the "date field restrictions" could be eased up to allow only valid
dates
> (as it is now) OR zero in a date field, this would help us a lot.
>
> I am currently having to using a packed field to store dates on the
> db, primarily due to this restriction.  Oftentimes, I need to allow a
> zero in
a
> date field, and current restrictions will not allow this.


while I understand the wish, it is not likely to ever be changed. And I
think "help us" is a key mentality here... it sure won't help the developers
that follow us.

One should never use a value to represent a "no value" condition and that's
what you are doing. A null value is what is used to represent that
condition. Since nulls are viable in date fields, you would/should convert
your dates and null those dates that are zero.

And, yes, application code would all have to change then too.

===========================================================
R. Bruce Hoffman, Jr.