[tz] Trivial questions about Tz

dpatte dpatte at relativedata.com
Mon Aug 15 22:56:13 UTC 2022

Thanks Paul

It is theoretically possible that such a triple midnight could happen, but I assume it could happen only if some goverment/administration had poorly chosen their walltime transition times.

----- Original Message -----
 From: Paul Eggert [mailto:eggert at cs.ucla.edu]
 To: "dpatte" <dpatte at relativedata.com>
 Cc: tz at iana.org
 Sent: Mon, 15 Aug 2022 15:46:55 -0700
 Subject: Re: [tz] Trivial questions about Tz

 On 8/15/22 14:23, dpatte via tz wrote:
> Are there any known cases in the tz database where two close offset and save changes in one region cause wall times in that region to have not just one but two repeats of the original wall times, such as 3 midnights with the same date?

I dimly recall we used to have examples of that but eventually removed 
them from TZDB since they were surely mistakes. Quite possibly my memory 
is incorrect.

> I am also curious what the largest single change to a save-time is in the database, and the largest single change to a GMT offset-time (though I assume the answer to this second one is Western Samoa).

The following shell command:

  zdump -i $(awk '/^Zone/{print $2}' main.zi) |
  awk 'o!="" && $3!="" && o!=$3 {print o, $3} {o = $3}' |
  sort -u

prints output that should help answer your question. Each line of output 
is a distinct pair of (before, after) %z-style abbreviations that is 
observed somewhere in tzdb. I get 579 lines of output with current tzdb.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/tz/attachments/20220815/3ecc1c2a/attachment.htm>

More information about the tz mailing list