Should NoneType be iterable?

[Peter Bona]

Hi

I am wondering if there has been any discussion why NoneType is not iterable My feeling is that it should be.
Sometimes I am using API calls which return None.
If there is a return value (which is iterable) I am using a for loop to iterate.

Now I am getting 'TypeError: 'NoneType' object is not iterable'.

(Examples are taken from here https://rollbar.com/blog/python-typeerror-nonetype-object-is-not-iterable/)
Example 1:
mylist = None
for x in mylist:
print(x) <== will raise TypeError: 'NoneType' object is not iterable
Solution: extra If statement
if mylist is not None:
for x in mylist:
print(x)


I think Python should handle this case gracefully: if a code would iterate over None: it should not run any step. but proceed the next statement.

Has this been discussed or proposed?

Thanks
Peter

[Julio Di Egidio]

I don't know about that, but maybe (also) consider this:

```
for x in mylist or []:
print(x)
```

which is a rather common pattern with dynamic languages.

Julio

[Chris Angelico]

Try this instead:

for x in mylist or ():

Now a None list will skip iteration entirely, allowing you to get the
effect you want :)

ChrisA

[Julio Di Egidio]

I'd expect the one with the if in front to be the most efficient.

Yours, unless there is some magic behind the implementation
of "for" I don't know about, still becomes an iteration over an
empty iterable, i.e. the same as with "[]". No?

Or, are you implying that the conversion of () to an empty
iterable itself is more efficient than that of []? Or, are you
saying something else altogether?

Julio

[Neal Becker]

On Mon, Jun 19, 2023 at 12:42 PM Chris Angelico via Python-list <
pytho...@python.org> wrote:

> On Tue, 20 Jun 2023 at 02:37, Peter Bona via Python-list
> <pytho...@python.org> wrote:
> >

> Try this instead:
>
> for x in mylist or ():
>
> Now a None list will skip iteration entirely, allowing you to get the
> effect you want :)
>

> ChrisA
> --
> https://mail.python.org/mailman/listinfo/python-list
>
I prefer iteration of None to be an error, as in my usage it usually
indicates a mistake that I'd want to catch

[Julio Di Egidio]

> I prefer iteration of None to be an error, as in my usage it usually
> indicates a mistake that I'd want to catch

That is quite problematic in itself:
If None is truly invalid for mylist,
one should report failure, possibly throw,
immediately after receiving the invalid value.
Letting it fail later and eventually unpredictably,
meanS not only useless stack traces, but also,
and more importantly, that any intervening side
effects have been effected, with unpredictable
outcomes, potentially disastrous.

If None is a legitimate value for mylist, of course
don't just throw to mimic control flow.

Julio

[dn]

On 20/06/2023 06.12, Neal Becker via Python-list wrote:
> On Mon, Jun 19, 2023 at 12:42 PM Chris Angelico via Python-list <
> pytho...@python.org> wrote:
>
>> On Tue, 20 Jun 2023 at 02:37, Peter Bona via Python-list
>> <pytho...@python.org> wrote:
>>>

>> Try this instead:
>>
>> for x in mylist or ():
>>
>> Now a None list will skip iteration entirely, allowing you to get the
>> effect you want :)
>>

>> ChrisA
>> --
>> https://mail.python.org/mailman/listinfo/python-list

>>
> I prefer iteration of None to be an error, as in my usage it usually
> indicates a mistake that I'd want to catch

Agreed!

A better approach is that the API return (perhaps a tuple of) both
"status" and "return_value", rather than overloading the latter.

That said, apparently the OP use-case is for when there is no interest
in status/catch, eg where a 'nothing' answer is THE answer.

--
Regards,
=dn

[Julio Di Egidio]

On Monday, 19 June 2023 at 21:58:21 UTC+2, dn wrote:
> On 20/06/2023 06.12, Neal Becker via Python-list wrote:

> > I prefer iteration of None to be an error, as in my usage it usually
> > indicates a mistake that I'd want to catch
>
> Agreed!

That is a *bad practice* with potentially disastrous consequences,
as I have explained. You too, like Angelico, can't even read?

> A better approach is that the API return (perhaps a tuple of) both
> "status" and "return_value", rather than overloading the latter.

Which is some pattern for public API's, otherwise just raise an
error where an error is the case (so called "exceptions").

> That said, apparently the OP use-case is for when there is no interest
> in status/catch, eg where a 'nothing' answer is THE answer.

A answer, or the question would be moot to begin with.

Other than that, of course we agree.

Julio

[Igor Berger]

On Monday, June 19, 2023 at 4:11:15 PM UTC-4, Julio Di Egidio wrote:
> On Monday, 19 June 2023 at 21:58:21 UTC+2, dn wrote:
> > On 20/06/2023 06.12, Neal Becker via Python-list wrote:
>
> > > I prefer iteration of None to be an error, as in my usage it usually
> > > indicates a mistake that I'd want to catch
> >
> > Agreed!
> That is a *bad practice* with potentially disastrous consequences,
> as I have explained. You too, like Angelico, can't even read?
>

> [snip]
>
> Julio

Julio,

Most of the regulars in this list/group read the posts using the mailing list and
have declared that they explicitly filter out anything posted on Google Groups.

I've seen it multiple times with your posts. You respond to something and others post parallel to you.
It looks like you're being ignored. However, they simply don't see your posts.

Neither they'll see my response.

Regards,
Igor.

[Greg Ewing]

I would question the wisdom of designing an API that
can return either a sequence or None. If it normally
returns a sequence, and there are no items to return,
it should return an empty sequence.

--
Greg

[Barry]

> On 20 Jun 2023, at 01:57, Greg Ewing via Python-list <pytho...@python.org> wrote:
>
> I would question the wisdom of designing an API that
> can return either a sequence or None.

I have some APIs that do return None or a list.
The None says that a list is not available and that the caller is
responsible with dealing in a application-domain specific with
with that situation.

In other cases I have API that returns a default list, often [].

It all depends on the design needs.

Barry

[Julio Di Egidio]

On Tuesday, 20 June 2023 at 02:19:02 UTC+2, Igor Berger wrote:
> On Monday, June 19, 2023 at 4:11:15 PM UTC-4, Julio Di Egidio wrote:
> > On Monday, 19 June 2023 at 21:58:21 UTC+2, dn wrote:
> > > On 20/06/2023 06.12, Neal Becker via Python-list wrote:
> >
> > > > I prefer iteration of None to be an error, as in my usage it usually
> > > > indicates a mistake that I'd want to catch
> > >
> > > Agreed!
> >
> > That is a *bad practice* with potentially disastrous consequences,
> > as I have explained. You too, like Angelico, can't even read?
> >
> > [snip]
>

> Most of the regulars in this list/group read the posts using the mailing list and
> have declared that they explicitly filter out anything posted on Google Groups.
>
> I've seen it multiple times with your posts. You respond to something and others post parallel to you.
> It looks like you're being ignored. However, they simply don't see your posts.

No, these very same guys did see my posts initially and I bet they still do:
rather you don't discount the fact that there are trolls and spammers and
plain charlatans encroaching this as every community, and I have never
shied away from calling out these nazi-retarded enemies of life and
intelligence. (And YMMV about that, but then don't complain if I call you
part of the problem).

That said, I know some people don't read GG and I couldn't care less: I
write for those who can and do read, and too bad for all the others...

Julio

[Greg Ewing]

On 20/06/23 7:36 pm, Barry wrote:
> I have some APIs that do return None or a list.
> The None says that a list is not available and that the caller is
> responsible with dealing in a application-domain specific with
> with that situation.

In that case, the caller should probably be checking for
None rather than blindly trying to iterate over the result.

--
Greg

[Roel Schroeven]

Op 20/06/2023 om 2:50 schreef Greg Ewing via Python-list:

I guess it depends on the reason why there are no items. If it is simply
because there are no items, then yes, I agree it should return an empty
sequence. But if it is because of some kind of error condition, I don't
think it should. But in that case I don't think the API should return
None either: I feel it should raise an exception.

--
"Je ne suis pas d’accord avec ce que vous dites, mais je me battrai jusqu’à
la mort pour que vous ayez le droit de le dire."
-- Attribué à Voltaire
"I disapprove of what you say, but I will defend to the death your right to
say it."
-- Attributed to Voltaire
"Ik ben het niet eens met wat je zegt, maar ik zal je recht om het te zeggen
tot de dood toe verdedigen"
-- Toegeschreven aan Voltaire

[Chris Angelico]

On Wed, 21 Jun 2023 at 03:46, Igor Berger via Python-list
<pytho...@python.org> wrote:
> Most of the regulars in this list/group read the posts using the mailing list and
> have declared that they explicitly filter out anything posted on Google Groups.
>
> I've seen it multiple times with your posts. You respond to something and others post parallel to you.
> It looks like you're being ignored. However, they simply don't see your posts.
>
> Neither they'll see my response.
>

I don't filter out Google Groups. However, the mailing list admins
have banned certain users, and their posts do not make it across the
gateway. So for those particular people, NONE of us see their posts.

If you post on Google Groups and it seems like almost nobody is
reading your posts, check whether you've been consistently violating
the Python Code of Conduct. Maybe the problem isn't with everyone
else.

ChrisA

[Cameron Simpson]

I wasted some time the other evening on an API which returned a string
or None. My own API, and the pain it caused tells me that that API
design choice isn't good (it's an automatic attribute based on a tag,
returning None if the tag isn't there). My experience so far is that it
_looks_ handy so that you can safely say "foo.bar" all the time, but as
soon a you do something with the value you're in for a world of
None-checking.

I'm rethinking that choice right now. Just the other day I removed a
setting in a related class which provided an automatic default value
because, again, while handy for careless use it caused hard to debug
problems because the default would flow out the call chain until it was
unsuitable, making the cause hard to trace.

And of course I'm very -1 on None acquiring iteration or other features.
Fail early, fail often!

Cheers,
Cameron Simpson <c...@cskk.id.au>

[Chris Angelico]

On Wed, 21 Jun 2023 at 09:59, Cameron Simpson via Python-list
<pytho...@python.org> wrote:

> I wasted some time the other evening on an API which returned a string
> or None. My own API, and the pain it caused tells me that that API
> design choice isn't good (it's an automatic attribute based on a tag,
> returning None if the tag isn't there). My experience so far is that it
> _looks_ handy so that you can safely say "foo.bar" all the time, but as
> soon a you do something with the value you're in for a world of
> None-checking.
>

https://peps.python.org/pep-0505/

ChrisA

[Cameron Simpson]

Hmm. Thanks. - Cameron Simpson <c...@cskk.id.au>