On Jul 9, 6:00 pm, Karl & Betty Schendel <schen...@kbcomputer.com>
wrote:
> At 12:50 AM +0000 7/9/07, droesler wrote:
>
> >In 6.4 ingres we could use the TZ env variable to determine what TZ
> >offset to apply to date types with time in them when displayed. This
> >doesn't seem to be the case in 2.6 or 3.x.
>
> >I can change II_TIMEZONE_NAME on the fly and it will use that, but it
> >seems that someone once said "don't do that". Does ingres no longer
> >use the TZ variable and is it OK to change II_TIMEZONE_NAME on the
> >fly?

> Ingres does not use TZ, and it should be OK to change II_TIMEZONE_NAME
> "on the fly" (meaning at session connect) as long as you understand
> the implications.
Thanks. I can't remember who said "don't do that".

> >Somewhat offf-post, but is there a reason Ingres doesn't use the
> >standard timezone names for their timezone names? For example,
> >America/New_York instead of NA-EASTERN?
>
> Are there in fact standard timezone names?

Maybe standard wasn't the term to use. I was referring to the
zoneinfo database.
Possibly GMT[+|-]nn may be the most reliable method across systems,
but that doesn't seem to suggest DST if it is applicable.

> Part of the reason for
> the entire II_TIMEZONE_NAME business is that there wasn't any
> reliably portable library level timezone support, at least not in
> the mid-90's when the code was done. Naming was the least of
> their problems. Even now I'm not aware of a world standard
> for timezone names. (not that I have looked very hard...)

Not that wikipedia is always the definitive answer :-), but there is
this
http://en.wikipedia.org/wiki/Zoneinfo

As always, appreciate your input Karl.

Dennis