Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #64640 > unrolled thread

Re: datetime as subclass of date

Back to article view | Back to comp.lang.python

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: datetime as subclass of date Ben Finney <ben+python@benfinney.id.au> - 2014-01-24 09:14 +1100
    Re: datetime as subclass of date Roy Smith <roy@panix.com> - 2014-01-23 19:10 -0500
      Re: datetime as subclass of date Ben Finney <ben+python@benfinney.id.au> - 2014-01-24 17:54 +1100
      Re: datetime as subclass of date Skip Montanaro <skip@pobox.com> - 2014-01-24 10:03 -0600

#64640 — Re: datetime as subclass of date

Skip Montanaro <skip@pobox.com> writes:

> […] I was asking [Python] if a datetime instance was an instance of a
> date. Which, it turns out, it is.

Yep. Makes sense, since ‘datetime’ can do everything ‘date’ can do, and
is conceptually a subset of the same concept.

The same is not true of the ‘time’ type from that module, by the way; it
does not represent a specific point on the timeline. So ‘datetime’
inheriting from ‘time’ wouldn't make sense. Just in case anyone was
wondering.

-- 
 \     “There is something wonderful in seeing a wrong-headed majority |
  `\           assailed by truth.” —John Kenneth Galbraith, 1989-07-28 |
_o__)                                                                  |
Ben Finney

[toc] | [next] | [standalone]

#64643

In article <mailman.5913.1390515296.18130.python-list@python.org>,
 Ben Finney <ben+python@benfinney.id.au> wrote:

> Skip Montanaro <skip@pobox.com> writes:
> 
> > […] I was asking [Python] if a datetime instance was an instance of a
> > date. Which, it turns out, it is.
> 
> Yep. Makes sense, since ‘datetime’ can do everything ‘date’ can do, and
> is conceptually a subset of the same concept.

That's reasonable, but given that, it's weird that date(2014, 1, 23) == 
datetime(2014, 1, 23) is False.  You would think it should be True, in 
the same way that 1 + 0j == 1 is True.

[toc] | [prev] | [next] | [standalone]

#64663

Roy Smith <roy@panix.com> writes:

> Ben Finney <ben+python@benfinney.id.au> wrote:
>
> > Makes sense, since ‘datetime’ can do everything ‘date’ can do, and
> > is conceptually a subset of the same concept.
>
> That's reasonable, but given that, it's weird that date(2014, 1, 23) == 
> datetime(2014, 1, 23) is False.  You would think it should be True, in 
> the same way that 1 + 0j == 1 is True.

Hmm. It does make sense to me that ‘datetime.date(2014, 1, 23) ==
datetime.datetime(2014, 1, 23)’.

I can come up with rationalisations for why it isn't, but they're not
satisfactory. I also don't consider it a bug, though. I'm conflicted :-)

-- 
 \          “Those who write software only for pay should go hurt some |
  `\                 other field.” —Erik Naggum, in _gnu.misc.discuss_ |
_o__)                                                                  |
Ben Finney

[toc] | [prev] | [next] | [standalone]

#64691

One thing that always reinforced my notion that
issubclass(datetime.datetime, datetime.date) should be False is that
the presence of of date and time methods gives me a mental image of
delegation, not inheritance. That is, it "feels" like a datetime
object is the aggregation of a date object and a time object, not a
specialized date object with some added time machinery mixed in.

Skip

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.python

csiph-web