Re: different date/time representations

[Mike Miller]

On Thu, May 26, 2016 at 12:17:33 +0200, Andreas Weber wrote:
> Dear all,
> 
> Since I'm always baffled because the different representations for
> date/time in GNU Octave (serial day number, date vector, time structure
> and so on) I made a graphic to help me when dealing with date/time:
> 
> http://www.tech-chat.de/images/GNU_Octave_date_time.svg

Nice chart! Could use some color :)

> Am I missing something? Should I add this to the wiki?
> 
> And is there a direct way from "seconds since the epoch" to "serial day
> number"?
> A function like:
> 
> tm = localtime (t);
> ret = datenum (tm.year + 1900, tm.mon + 1, tm.mday, tm.hour, tm.min,
> tm.sec + tm.usec / 1e6)

Sure, this conversion works for me directly from a time_t:

  utc_offset = -7;
  num = time () / 86400 + utc_offset / 24 + 719529;

You could also use the gmtoff field:

  t = time ();
  num = (t + localtime (t).gmtoff) / 86400 + 719529;

I found 719529 empirically from

  datenum (1970, 1, 1, 0, 0, 0)

presumably the number of days from 0000-01-01 to 1970-01-01.

-- 
mike

signature.asc
Description: PGP signature