python-peps/pep-0495.txt

PEP: 495
Title: Local Time Disambiguation
Version: $Revision$
Last-Modified: $Date$
Author: Alexander Belopolsky <alexander.belopolsky@gmail.com>
Discussions-To: Datetime-SIG <datetime-sig@python.org>
Status: Draft
Type: Standards Track
Content-Type: text/x-rst
Created: 02-Aug-2015



Abstract
========

This PEP adds a boolean member to the instances of ``datetime.time``
and ``datetime.datetime`` classes that can be used to differentiate
between two moments in time for which local times are the same.

.. sidebar:: US public service advertisement

  .. image:: pep-0495-daylightsavings.png
     :align: right
     :width: 15%


Rationale
=========

In the most world locations there have been and will be times when
local clocks are moved back.  In those times, intervals are introduced
in which local clocks show the same time twice in the same day.  In
these situations, the information displayed on a local clock (or
stored in a Python datetime instance) is insufficient to identify a
particular moment in time.  The proposed solution is to add a boolean
flag to the ``datetime`` instances that will distinguish between the
two ambiguous times.


Terminology
===========

When clocks are moved back, we say that a *fold* is created in the
fabric of time.  When the clock are moved forward, a *gap* is created.
A local time that falls in the fold is called *ambiguous*.  A local
time that falls in the gap is called *missing*.


Proposal
========

The "first" flag
----------------

We propose adding a boolean member called ``first`` to the instances
of ``datetime.time`` and ``datetime.datetime`` classes.  This member
should have the value True for all instances except those that
represent the second (chronologically) moment in time in an ambiguous
case. [#]_

.. [#] An instance that has ``first=False`` in a non-ambiguous case is
  said to represent an invalid time (or is invalid for short), but
  users are not prevented from creating invalid instances by passing
  ``first=False`` to a constructor or to a ``replace()`` method.  This
  is similar to the current situation with the instances that fall in
  the spring-forward gap.  Such instances don't represent any valid
  time, but neither the constructors nor the ``replace()`` methods
  check whether the instances that they produce are valid.  Moreover,
  this PEP specifies how various functions should behave when given an
  invalid instance.

Affected APIs
-------------

Attributes
..........

Instances of ``datetime.time`` and ``datetime.datetime`` will get a
new boolean attribute called "first."

Constructors
............

The ``__new__`` methods of the ``datetime.time`` and
``datetime.datetime`` classes will get a new keyword-only argument
called ``first`` with the default value ``True``.  The value of the
``first`` argument will be used to initialize the value of the
``first`` attribute in the returned instance.

Methods
.......

The ``replace()`` methods of the ``datetime.time`` and
``datetime.datetime`` classes will get a new keyword-only argument
called ``first``.  It will
behave similarly to the other ``replace()`` arguments: if the ``first``
argument is specified and given a boolean value, the new instance
returned by ``replace()`` will have its ``first`` attribute set
to that value.  In CPython, a non-boolean value of ``first`` will
raise a ``TypeError``, but other implementations may allow the value
``None`` to behave the same as when ``first`` is not given.  If the
``first`` argument is not specified, the original value of the ``first``
attribute is copied to the result.   

C-API
.....

Access macros will be defined to extract the value of ``first`` from
``PyDateTime_DateTime`` and ``PyDateTime_Time`` objects.

.. code:: 

  bool PyDateTime_GET_FIRST(PyDateTime_DateTime *o)

Return the value of ``first`` as a C99 ``bool``.

.. code:: 
  
  bool PyDateTime_TIME_GET_FIRST(PyDateTime_Time *o)

Return the value of ``first`` as a C99 ``bool``.

Additional constructors will be defined that will take an additional
boolean argument to specify the value of ``first`` in the created
instance:

.. code::

  PyObject* PyDateTime_FromDateAndTimeAndFirst(int year, int month, int day, int hour, int minute, int second, int usecond, bool first)

Return a ``datetime.datetime`` object with the specified year, month,
day, hour, minute, second, microsecond and first.

.. code::

  PyObject* PyTime_FromTimeAndFirst(int hour, int minute, int second, int usecond, bool first)

Return a ``datetime.time`` object with the specified hour, minute,
second, microsecond and first.


Affected Behaviors
------------------

Conversion from naive to aware
..............................

The ``astimezone()`` method will now work for naive ``self``.  The
system local timezone will be assumed in this case and the ``first``
flag will be used to determine which local timezone is in effect
in the ambiguous case.

For example, on a system set to US/Eastern timezone::

  >>> dt = datetime(2014, 11, 2, 1, 30)
  >>> dt.astimezone().strftime('%D %T %Z%z')
  '11/02/14 01:30:00 EDT-0400'
  >>> dt.replace(first=False).astimezone().strftime('%D %T %Z%z')
  '11/02/14 01:30:00 EST-0500'

Conversion to POSIX seconds from EPOCH
......................................

The ``timestamp()`` method of ``datetime.datetime`` will return different
values for ``datetime.datetime`` instances that differ only by the value
of their ``first`` attribute if and only if these instances represent an
ambiguous or a non-existent time.

When a ``datetime.datetime`` instance ``dt`` represents an ambiguous
(repeated) time, there are two values ``s0`` and ``s1`` such that::

  datetime.fromtimestamp(s0) == datetime.fromtimestamp(s1) == dt

In this case, ``dt.timestamp()`` will return the smaller of ``s0``
and ``s1`` values if ``dt.first == True`` and the larger otherwise.


For example, on a system set to US/Eastern timezone::

  >>> datetime(2014, 11, 2, 1, 30, first=True).timestamp()
  1414906200.0
  >>> datetime(2014, 11, 2, 1, 30, first=False).timestamp()
  1414909800.0


When a ``datetime.datetime`` instance ``dt`` represents a missing
time, there is no value ``s`` for which::

  datetime.fromtimestamp(s) == dt

but we can form two "nice to know" values of ``s`` that differ
by the size of the gap in seconds.  One is the value of ``s``
that would correspond to ``dt`` in a timezone where the UTC offset
is always the same as the offset right before the gap and the
other is the similar value but in a timezone the  UTC offset
is always the same as the offset right after the gap.

The value returned by ``dt.timestamp()`` given a missing
``dt`` will be the larger of the two "nice to know" values
if ``dt.first  == True`` and the larger otherwise.

For example, on a system set to US/Eastern timezone::

  >>> datetime(2015, 3, 8, 2, 30, first=True).timestamp()
  1425799800.0
  >>> datetime(2015, 3, 8, 2, 30, first=False).timestamp()
  1425796200.0


Conversion from POSIX seconds from EPOCH
........................................


The ``fromtimestamp()`` static method of ``datetime.datetime`` will
set the ``first`` attribute appropriately in the returned object.

For example, on a system set to US/Eastern timezone::

  >>> datetime.fromtimestamp(1414906200)
  datetime.datetime(2014, 11, 2, 1, 30)
  >>> datetime.fromtimestamp(1414906200 + 3600)
  datetime.datetime(2014, 11, 2, 1, 30, first=False)


Combining and splitting date and time
.....................................

The ``datetime.datetime.combine()`` method will copy the value of the
``first`` attribute to the resulting ``datetime.datetime`` instance.

The ``datetime.datetime.time()`` method will copy the value of the
``first`` attribute to the resulting ``datetime.time`` instance.


Implementations of tzinfo in stdlib
...................................

No new implementations of ``datetime.tzinfo`` abstract class are
introduced in this PEP.  The existing (fixed offset) timezones do
not introduce ambiguous local times and their ``utcoffset()``
implementation will return the same constant value as they do now
regardless of the value of ``first``. 

The basic implementation of ``fromutc()`` in the abstract
``datetime.tzinfo`` class will not change.  It is currently not
used anywhere in the stdlib because the only included ``tzinfo``
implementation (the ``datetime.timzeone`` class implementing fixed
offset timezones) override ``fromutc()``.

New guidelines will be published for implementing concrete timezones
with variable UTC offset.


Guidelines for new tzinfo implementations
-----------------------------------------

Implementors of concrete ``datetime.tzinfo`` subclasses who want to
support variable UTC offsets (due to DST and other causes) must follow
these guidelines.

New subclasses must override the base-class ``fromutc()`` method and
implement it so that in all cases where two UTC times ``u1`` and
``u2`` (``u1`` <``u2``) correspond to the same local time
``fromutc(u1)`` will return an instance with ``first=True`` and
``fromutc(u2)`` will return an instance with ``first=False``.  In all
other cases the returned instance must have ``first=True``.

New implementations of ``utcoffset()`` and ``dst()`` methods should
ignore the value of ``first`` unless they are called on the ambiguous
or missing times.

On an ambiguous time introduced at the end of DST, the values returned
by ``utcoffset()`` and ``dst()`` methods should be as follows

+-----------------+----------------+------------------+
|                 |   first=True   |    first=False   |
+=================+================+==================+
|   utcoff()      | stdoff + hour  |    stdoff        |
+-----------------+----------------+------------------+
|    dst()        |     hour       |     zero         |
+-----------------+----------------+------------------+

where ``stdoff`` is the standard (non-DST) offset,
``hour = timedelta(hours=1)`` and ``zero = timedelta(0)``.

On a missing time introduced at the start of DST, the values returned
by ``utcoffset()`` and ``dst()`` methods should	be as follows


+-----------------+----------------+------------------+
|                 |   first=True   |    first=False   |
+=================+================+==================+
|   utcoff()      |     stdoff     |    stdoff + hour |
+-----------------+----------------+------------------+
|    dst()        |     zero       |     hour         |
+-----------------+----------------+------------------+


On ambiguous/missing times introduced by the change in the standard time
offset, the ``dst()`` method should return the same value regardless of
the value of ``first`` and the ``utcoff()`` should return values
according to the following table:

+-----------------+----------------+-----------------------------+
|                 |   first=True   |    first=False              |
+=================+================+=============================+
|   ambiguous     |     oldoff     |     newoff = oldoff - delta |
+-----------------+----------------+-----------------------------+
|   missing       |     oldoff     |     newoff = oldoff + delta |
+-----------------+----------------+-----------------------------+



Pickle size
-----------

Pickle sizes for the ``datetime.datetime`` and ``datetime.time``
objects will not change.  The ``first`` flag will be encoded in the
first bit of the 5th byte of the ``datetime.datetime`` pickle payload
or the 2nd byte of the datetime.time. In the `current implementation`_
these bytes are used to store minute value (0-59) and the first bit is
always 0.  Note that ``first=True`` will be encoded as 0 in the first
bit and ``first=False`` as 1.  (This change only affects pickle
format.  In C implementation, the "first" member will get a full byte
to store the actual boolean value.)

We chose the minute byte to store the the "first" bit because this
choice preserves the natural ordering.

.. _current implementation: https://hg.python.org/cpython/file/d3b20bff9c5d/Include/datetime.h#l17

Temporal Arithmetics
--------------------

The value of "first" will be ignored in all operations except those
that involve conversion between timezones. [#]_ As a consequence,
`datetime.datetime`` or ``datetime.time`` instances that differ only
by the value of ``first`` will compare as equal.  Applications that
need to differentiate between such instances should check the value of
``first`` or convert them to a timezone that does not have ambiguous
times.

The result of addition (subtraction) of a timedelta to (from) a
datetime will always have ``first`` set to ``True`` even if the
original datetime instance had ``first=False``.

.. [#] As of Python 3.5, ``tzinfo`` is ignored whenever timedelta is
   added or subtracted from a ``datetime.datetime`` instance or when
   one ``datetime.datetime`` instance is subtracted from another with
   the same (even not-None) ``tzinfo``.  This may change in the future,
   but such changes are outside of the scope of this PEP.


Backward and Forward Compatibility
----------------------------------

This proposal will have little effect on the programs that do not read
the ``first`` flag explicitly or use tzinfo implementations that do.
The only visible change for such programs will be that conversions to
and from POSIX timestamps will now round-trip correctly (up to
floating point rounding).  Programs that implemented work-arounds to
the old incorrect behavior will need to be modified.

Pickles produced by older programs will remain fully forward
compatible.  Only datetime/time instances with ``first=False`` pickled
in the new versions will become unreadable by the older Python
versions.  Pickles of instances with ``first=True`` (which is the
default) will remain unchanged.

Questions and Answers
=====================

1. Why not call the new flag "isdst"?

-------

* Alice:  Bob - let's have a stargazing party at 01:30 AM tomorrow!
* Bob:  Should I presume initially that summer time (for example, Daylight
  Saving Time) is or is not (respectively) in effect for the specified time?
* Alice: Huh?

-------

* Bob: Alice - let's have a stargazing party at 01:30 AM tomorrow!
* Alice: You know, Bob, 01:30 AM will happen twice tomorrow. Which time do you have in mind?
* Bob:  I did not think about it, but let's pick the first.


2. Why "first"?

* Rejections:

  **second**   
      rejected because "second" is already there.
   
  **later**
      rejected because "later" is confusable with "latter".
   
  **earlier**
      rejected because "earlier" has the same issue as "first" (requires
      default to be True) but is two characters longer.
   
* Remaining possibilities:
 
  **repeated**
      this is a strong candidate
 
  **is_first**
      arguably more grammatically correct than "first"
 
  **ltdf**
      (Local Time Disambiguation Flag) - short and no-one will
      attempt to guess what it means without reading the docs.

Implementation
==============

* Github fork: https://github.com/abalkin/cpython
* Tracker issue: http://bugs.python.org/issue24773


Copyright
=========

This document has been placed in the public domain.


Picture Credit
==============

This image is a work of a U.S. military or Department of Defense
employee, taken or made as part of that person's official duties. As a
work of the U.S. federal government, the image is in the public
domain.