Time_Zone-aware Day_of_Week ?

Time_Zone-aware Day_of_Week ?

[Natasha Kerensikova]

Hello,

I was implementing some human time formatting with the help of Ada 2005
standard package Ada.Calendar.Formatting, and realized I have an issue
with the function Day_of_Week with regards to time zone.

Everything else in Ada.Calendar.Formatting seems to support some
user-defined time offset, but not Day_of_Week. Is there a particular
reason for that?

Moreover, the description in the ARM of the function is quite terse:
"Returns the day of the week for Time. This is based on the Year, Month,
and Day values of Time."

However, all of Year, Month and Day functions do have the Time_Zone
argument that Day_of_Week lacks. So which time offset is Day_of_Week
supposed to use?

Prying into GNAT's implementation, it seems to use standard system
timezone, but can it be deduced from the ARM?

Is there a workaround to still get a Day_of_Week result that is
consistent with a user-given time offset?

I'm currently leaning towards using Ada.Calendar.Arithmetic.Difference
with a reference date, at midnight in the request time zone, but that
feels a bit inelegant. Is there a better way of doing it?


Thanks for your help,
Natasha

Re: Time_Zone-aware Day_of_Week ?

[Jeffrey R. Carter]

On 07/13/2015 06:26 AM, Natasha Kerensikova wrote:
> 
> Moreover, the description in the ARM of the function is quite terse:
> "Returns the day of the week for Time. This is based on the Year, Month,
> and Day values of Time."

Presumably the Year, Month, and Day returned by Ada.Calendar.Split.

> However, all of Year, Month and Day functions do have the Time_Zone
> argument that Day_of_Week lacks. So which time offset is Day_of_Week
> supposed to use?

I'd guess the time offset given by Ada.Calendar.Time_Zones.UTC_Time_Offset.

PragmARC.Date_Handler.Day_Of_Week gives you the day of week of any Year, Month,
and Day CE you care to give it, even for values of Year outside
Ada.Calendar.Year_Number. It can also give you the day of week for any
Ada.Calendar.Time, using the Year, Month, and Day returned by Ada.Calendar.Split.

The PragmAda Reusable Components are at

https://pragmada.x10hosting.com/pragmarc.htm

-- 
Jeff Carter
"He that hath no beard is less than a man."
Much Ado About Nothing
132

Re: Time_Zone-aware Day_of_Week ?

[brbarkstrom]

On Monday, July 13, 2015 at 12:45:48 PM UTC-4, Jeffrey R. Carter wrote:
> On 07/13/2015 06:26 AM, Natasha Kerensikova wrote:
> > 
> > Moreover, the description in the ARM of the function is quite terse:
> > "Returns the day of the week for Time. This is based on the Year, Month,
> > and Day values of Time."
> 
> Presumably the Year, Month, and Day returned by Ada.Calendar.Split.
> 
> > However, all of Year, Month and Day functions do have the Time_Zone
> > argument that Day_of_Week lacks. So which time offset is Day_of_Week
> > supposed to use?
> 
> I'd guess the time offset given by Ada.Calendar.Time_Zones.UTC_Time_Offset.
> 
> PragmARC.Date_Handler.Day_Of_Week gives you the day of week of any Year, Month,
> and Day CE you care to give it, even for values of Year outside
> Ada.Calendar.Year_Number. It can also give you the day of week for any
> Ada.Calendar.Time, using the Year, Month, and Day returned by Ada.Calendar.Split.
> 
> The PragmAda Reusable Components are at
> 
> https://pragmada.x10hosting.com/pragmarc.htm
> 
> -- 
> Jeff Carter
> "He that hath no beard is less than a man."
> Much Ado About Nothing
> 132

Depending on the precision you need, it might be worth thinking about
the difference between Civil Time (which uses Year, Month, Day, Hour,
Minute, and Second) and International Atomic Time (TAI - due to French
Standards).  TAI tracks seconds and is directly related to
GPS Time and one or two other variants.  The standard reference for
these different time systems is Seidelmann, P. K., 2006: Explanatory
Supplement to the Astronomical Almanac, University Science Books, 
Sausalito, CA.  If you need to calculate time differences frequently
and want to avoid the unpleasantries of having to deal with errors 
associated with time zone inputs (or lack thereof), TAI is probably
preferable.  This may be important for Fourier analysis of time series,
where the incomensurability of Civil Time can introduces errors into the
frequency spectrum derived from the data.  One satellite mission had to
expend a great deal of effort to find that the lack of a common time basis
caused some serious errors in their measurements of the Earth's reflection.

Bruce B.

Re: Time_Zone-aware Day_of_Week ?

[Vadim Godunko]

You can try Matreshka, it has some timezone and time formatting support.

http://forge.ada-ru.org/matreshka