[Numpy-discussion] timezones and datetime64

[Jonathan T. Niehof]

On 04/04/2013 03:03 AM, Daniele Nicolodi wrote:
> I'm not aware of any library that handles the conversion from UTC to
> TAI. I would like to know if there is one.

The CDF library does, although that's rather a sledgehammer to drive a 
thumbtack.

> I think that generally the issue is not relevant for any practical use
> of a timebase: there are not many applications requiring sub-second
> accuracy over many years periods.

Which issue? Leapseconds? Leap seconds are required for *second* 
accuracy. They're also required for time-tagging anything that takes 
place during a leapsecond. The lack of proper no-skip time support in 
Python datetimes has caused us headaches and it was annoying to find 
that datetime64 doesn't improve matters.

Keeping a leap second database up to date is annoying but not as bad as 
it could be: they're not arbitrary. Although they can occur monthly, 
they've only ever happened at June 30 and Dec 31, announced in January 
and July, respectively. So it would be easy to check the date of a 
leapsecond database and warn if 1) date we're processing is after June 
30 of a year AND 2) LSDB older than January same year (similar checks 
for the Dec. 31 opportunity.)

Er, I guess I'm volunteering to help :)

-- 
Jonathan Niehof
ISR-3 Space Data Systems
Los Alamos National Laboratory
MS-D466
Los Alamos, NM 87545

Phone: 505-667-9595
email: jniehof at lanl.gov

Correspondence /
Technical data or Software Publicly Available