«

Why does the following result in an IndexError?
I try to match an optional group, and then access it via its group
name. The group happens to not participate in the match, but is
obviously defined in the pattern.
The documentation says that, at least for numbered groups, "If a group
is contained in a part of the pattern that did not match, the
corresponding result is None.".
So I would expect None rather than an IndexError, which is (only?)
supposed to occur "If a string argument is not used as a group name in
the pattern".

I would expect named groups and numbered groups to be behave the same
way.
Where's my mistake?

>>> import re
>>> m = re.match('(?P<m>maybe)?yes', "yes")
>>> m.group(1)
>>> m.group('maybe')
Traceback (most recent call last):
File "<stdin>", line 1, in ?
IndexError: no such group

PS.: Please cc your reply to my e-mail address.

wrote:
> Why does the following result in an IndexError?
> I try to match an optional group, and then access it via its group
> name. The group happens to not participate in the match, but is
> obviously defined in the pattern.
>
>>>>m = re.match('(?P<m>maybe)?yes', "yes")

Uh, don't you need to actually *group* the first "yes" part, with
parentheses, to make it a group? Otherwise you aren't going to be
treating it as other than text-between-groups-that-can-be-skipped.

-Peter

david.reitter wrote:

> So I would expect None rather than an IndexError, which is (only?)
> supposed to occur "If a string argument is not used as a group name in
> the pattern".
That is exactly what does happen.

>
> I would expect named groups and numbered groups to be behave the same
> way.
> Where's my mistake?
Using 'maybe' as a group name when the only group in the pattern is called
'm'.

>
>>>> import re
>>>> m = re.match('(?P<m>maybe)?yes', "yes")
>>>> m.group(1)
>>>> m.group('maybe')
> Traceback (most recent call last):
> File "<stdin>", line 1, in ?
> IndexError: no such group
>

Peter Hansen wrote:
> wrote:
>
>> Why does the following result in an IndexError?
>> I try to match an optional group, and then access it via its group
>> name. The group happens to not participate in the match, but is
>> obviously defined in the pattern.
>>
>>>>> m = re.match('(?P<m>maybe)?yes', "yes")
>
> Uh, don't you need to actually *group* the first "yes" part, with
> parentheses, to make it a group? Otherwise you aren't going to be
> treating it as other than text-between-groups-that-can-be-skipped.

Oops... from Duncan's reply and a reread I can see I missed the point
entirely... I obviously thought you were expecting the "yes" to be in a
group too. Duh.

-Peter