ISO8601 classes provide a method to_*
to convert to its core equivalent:
ISO8601::DateTime#to_datetime # => #<DateTime: ...>
ISO8601::Date#to_date # => #<Date: ...>
ISO8601::Time#to_time # => #<Time: ...>
Core Date.parse
and DateTime.parse
don't allow reduced precision. For
example:
DateTime.parse('2014-05') # => ArgumentError: invalid date
But the standard covers this situation assuming any missing token as its lower value:
ISO8601::DateTime.new('2014-05').to_s # => "2014-05-01T00:00:00+00:00"
ISO8601::DateTime.new('2014').to_s # => "2014-01-01T00:00:00+00:00"
The same assumption happens in core classes with .new
:
DateTime.new(2014,5) # => #<DateTime: 2014-05-01T00:00:00+00:00 ((2456779j,0s,0n),+0s,2299161j)>
DateTime.new(2014) # => #<DateTime: 2014-01-01T00:00:00+00:00 ((2456659j,0s,0n),+0s,2299161j)>
Unmatched precison is handled strongly. Notice the time fragment is lost in
DateTime.parse
with no warning only if the loose precision is in the time
component.
ISO8601::DateTime.new('2014-05-06T101112') # => ISO8601::Errors::UnknownPattern
DateTime.parse('2014-05-06T101112') # => #<DateTime: 2014-05-06T00:00:00+00:00 ((2456784j,0s,0n),+0s,2299161j)>
ISO8601::DateTime.new('20140506T10:11:12') # => ISO8601::Errors::UnknownPattern
DateTime.parse('20140506T10:11:12') # => #<DateTime: 2014-05-06T10:11:12+00:00 ((2456784j,0s,0n),+0s,2299161j)>
Week dates raise an error when two digit days provied instead of return monday:
ISO8601::DateTime.new('2014-W15-02') # => ISO8601::Errors::UnknownPattern
DateTime.new('2014-W15-02') # => #<Date: 2014-04-07 ((2456755j,0s,0n),+0s,2299161j)>
DateTime#to_a
allows decomposing to an array of atoms:
atoms = ISO8601::DateTime.new('2014-05-31T10:11:12Z').to_a # => [2014, 5, 31, 10, 11, 12, '+00:00']
dt = DateTime.new(*atoms)
Ordinal dates keep the sign. 2014-001
is not the same as -2014-001
.
Fractional seconds for ISO8601::DateTime
and ISO8601::Time
are rounded to
one decimal.
ISO8601::DateTime.new('2015-02-03T10:11:12.12').second #=> 12.1
ISO8601::Time.new('T10:11:12.16').second #=> 12.2