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

Groups > comp.lang.python > #75407 > unrolled thread

converting ISO8601 date and time string representations to datetime

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

Contents

  converting ISO8601 date and time string representations to datetime Wolfgang Maier <wolfgang.maier@biologie.uni-freiburg.de> - 2014-07-31 11:11 +0200
    Re: converting ISO8601 date and time string representations to datetime Roy Smith <roy@panix.com> - 2014-07-31 07:52 -0400
      Re: converting ISO8601 date and time string representations to datetime Skip Montanaro <skip@pobox.com> - 2014-07-31 07:59 -0500
      Re: converting ISO8601 date and time string representations to datetime Roy Smith <roy@panix.com> - 2014-07-31 09:13 -0400
      Re: converting ISO8601 date and time string representations to datetime Wolfgang Maier <wolfgang.maier@biologie.uni-freiburg.de> - 2014-07-31 15:29 +0200

#75407 — converting ISO8601 date and time string representations to datetime

Hi,
I'm trying to convert ISO8601-compliant strings representing dates or 
dates and times into datetime.datetime objects.
I tried using the strptime method, but the problem here is that I can 
only specify one format argument, which can be used to parse either a 
full date/time string or just a date string, but not both.

While I could just try parsing with one format, then catch the 
ValueError in case it fails and try another format, it feels like there 
should be a more elegant way to do this.

I do know about the dateutil package, but I'd prefer to stick to the 
stdlib for this.

Any suggestions ?

Wolfgang

[toc] | [next] | [standalone]

#75410

In article <mailman.12461.1406797909.18130.python-list@python.org>,
 Wolfgang Maier <wolfgang.maier@biologie.uni-freiburg.de> wrote:

> Hi,
> I'm trying to convert ISO8601-compliant strings representing dates or 
> dates and times into datetime.datetime objects.

https://pypi.python.org/pypi/iso8601

or

https://pypi.python.org/pypi/isodate

There's also....
 
> I do know about the dateutil package, but I'd prefer to stick to the 
> stdlib for this.

Sadly, the stdlib datetime really doesn't make life easy for dealing 
with ISO-8601.  Dateutil is the classic answer, but it's slow. The 
isodate module cited above is what we use.

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

#75412

On Thu, Jul 31, 2014 at 6:52 AM, Roy Smith <roy@panix.com> wrote:
> Sadly, the stdlib datetime really doesn't make life easy for dealing
> with ISO-8601.  Dateutil is the classic answer, but it's slow.

A useful feature for dateutil would be a "sniff" function which, given
a date string that dateutil.parser.parse can handle, instead returns a
format string suitable for use with time.strptime.

Skip

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

#75414

On Jul 31, 2014, at 8:59 AM, Skip Montanaro wrote:

> On Thu, Jul 31, 2014 at 6:52 AM, Roy Smith <roy@panix.com> wrote:
>> Sadly, the stdlib datetime really doesn't make life easy for dealing
>> with ISO-8601.  Dateutil is the classic answer, but it's slow.
> 
> A useful feature for dateutil would be a "sniff" function which, given
> a date string that dateutil.parser.parse can handle, instead returns a
> format string suitable for use with time.strptime.

+0.8.  But, even more than that, dateutil should come with a library of strings for common formats, and a built-in function to parse 8601.  It is, after all, the international standard.


--
Roy Smith
roy@panix.com

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

#75417

On 31.07.2014 15:13, Roy Smith wrote:
> On Jul 31, 2014, at 8:59 AM, Skip Montanaro wrote:
>
>> On Thu, Jul 31, 2014 at 6:52 AM, Roy Smith <roy@panix.com> wrote:
>>> Sadly, the stdlib datetime really doesn't make life easy for dealing
>>> with ISO-8601.  Dateutil is the classic answer, but it's slow.
>>
>> A useful feature for dateutil would be a "sniff" function which, given
>> a date string that dateutil.parser.parse can handle, instead returns a
>> format string suitable for use with time.strptime.
>
> +0.8.  But, even more than that, dateutil should come with a library of strings for common formats, and a built-in function to parse 8601.  It is, after all, the international standard.
>

Thanks for your feedback.
I'll have a look at isodate as you suggested, Roy.

What this makes me wonder is whether there shouldn't be a builtin method 
for parsing ISO-8601 strings in the stdlib datetime module, just as 
there is .isoformat() for the reverse action already.
Alternatively, the existing format argument to datetime.strptime() could 
be made optional and when absent, strptime() could try to parse ISO-8601 ?

Wolfgang

[toc] | [prev] | [standalone]

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

csiph-web