[Python-ideas] Re: Undefined type

[Dom Grigonis]

unittest.mock.sentinel did the trick for me. As long as there is something for 
it in a standard library for it I think it’s covered. Everyone defining their 
own classes or objects would be confusing.

> On 12 Jun 2023, at 14:27, Rob Cliffe via Python-ideas 
>  wrote:
> 
> The problem with adding a second None-like type is that gradually it will 
> acquire meaning as a possible value in some contexts.  Maybe not in your 
> code, but in other people's.  Then will there be a request for a 3rd such 
> object?  (4th, 5th ...?).
> Defining your own sentinel object, as you do, seems like a good solution.  It 
> means *you* control what meaning it has in the contexts where it appears.
> Best wishes
> Rob Cliffe
> 
> On 07/06/2023 17:43, Dom Grigonis wrote:
>> This has been bugging me for a long time. It seems that python doesn’t have 
>> a convention for “Undefined” type.
>> 
>> When I started using python I naturally used None for it (as advised).
>> 
>> However, the more I work with python, the more I find that None is a 
>> meaningful type. And it is a type that is convenient to use in a lot of 
>> cases, where it is not actually “None”.
>> 
>> One big example is:
>> numpy: array[None, :]
>> 
>> Another is simple cascading of defaults.
>> E.g.
>> 
>> ```python
>> class A:
>> def __init__(self, default=’string'):
>> self.default = default
>> 
>> def method(self, default=None):
>> default = default if default is not None else self.default
>>  dct[‘key’] = other_dict.get(‘key’, default)
>> return to_json(dct)
>> ```
>> 
>> None has a logic in `some_func`, which is important.
>> Now if I want to enforce None passed to to_json I have no way to do it.
>> 
>> I know there are many workarounds for this, but there is no `consistent` way 
>> of doing this.
>> 
>> I have been using Ellipsis in the past, but it doesn’t feel right, because 
>> it also has a meaning and I might need to use those functions with e.g. 
>> numpy in he future and will have to change it.
>> 
>> Now, I came back to this issue again and my latest solution is to just 
>> define:
>> UNDEFINED = object()
>> in library constants.py
>> 
>> Then code above is very clear and I think this is a good and sustainable 
>> solution.
>> 
>> I know, that in a way ’None’ is already this, but it is generally used in 2 
>> ways:
>> 1. A variable/argument is undefined in python space
>> 2. Global indication, that a value is undefined (e.g. equivalence of null in 
>> json)
>> 
>> What I am proposing is to have 1 more constant in python ‘Undefined’, which 
>> will indicate ‘Undefined’ in python space and leave None to be used in more 
>> general sense.
>> 
>> I also found this article while looking for solution just now:
>> https://levelup.gitconnected.com/python-why-none-is-not-nothing-bb3de55dd471 
>> 
>> It seems that people who use typing are also running into ambiguities and 
>> trying to hack something together.
>> I avoid typing until its 99% mature so can't comment more on this.
>> 
>> What are your thoughts? Am I onto something or am I missing something? 
>> Anyone else had similar encounters?
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> ___
>> Python-ideas mailing list -- python-ideas@python.org 
>> 
>> To unsubscribe send an email to python-ideas-le...@python.org 
>> 
>> https://mail.python.org/mailman3/lists/python-ideas.python.org/ 
>> 
>> Message archived at 
>> https://mail.python.org/archives/list/python-ideas@python.org/message/RGUNJYXWHE7UM2W6FCAMGLJVLXEYHA2Q/
>>  
>> 
>> Code of Conduct: http://python.org/psf/codeofconduct/ 
>> 
> 
> ___
> Python-ideas mailing list -- python-ideas@python.org
> To unsubscribe send an email to python-ideas-le...@python.org
> https://mail.python.org/mailman3/lists/python-ideas.python.org/
> Message archived at 
> https://mail.python.org/archives/list/python-ideas@python.org/message/2EH4RSB6LPZLV7OOQOOI65DWFSKCZNS2/
> Code of Conduct: http://python.org/psf/codeofconduct/

___
Python-ideas mailing list -- python-ideas@python.org
To unsubscribe send an email to python-ideas-le...@python.org
https://mail.python.org/mailman3/lists/python-ideas.python.org/
Message archived at 
https://mail.python.org/archives/list/python-ideas@python.org/message/QBLI2ZWOKUB7BEPAZSKAQ63T3WEWIR4M/
Code of Conduct: http://python.org/psf/codeofconduct/

[Python-ideas] Re: Undefined type

[Rob Cliffe via Python-ideas]

The problem with adding a second None-like type is that gradually it 
will acquire meaning as a possible value in some contexts.  Maybe not in 
your code, but in other people's.  Then will there be a request for a 
3rd such object?  (4th, 5th ...?).
Defining your own sentinel object, as you do, seems like a good 
solution.  It means *you* control what meaning it has in the contexts 
where it appears.

Best wishes
Rob Cliffe

On 07/06/2023 17:43, Dom Grigonis wrote:
This has been bugging me for a long time. It seems that python doesn’t 
have a convention for “Undefined” type.


When I started using python I naturally used None for it (as advised).

However, the more I work with python, the more I find that None is a 
meaningful type. And it is a type that is convenient to use in a lot 
of cases, where it is not actually “None”.


One big example is:
numpy: array[None, :]

Another is simple cascading of defaults.
E.g.

```python
class A:
    def __init__(self, default=’string'):
        self.default = default

    def method(self, default=None):
        default = default if default is not None else self.default
dct[‘key’] = other_dict.get(‘key’, default)
        return to_json(dct)
```

None has a logic in `some_func`, which is important.
Now if I want to enforce None passed to to_json I have no way to do it.

I know there are many workarounds for this, but there is no 
`consistent` way of doing this.


I have been using Ellipsis in the past, but it doesn’t feel right, 
because it also has a meaning and I might need to use those functions 
with e.g. numpy in he future and will have to change it.


Now, I came back to this issue again and my latest solution is to just 
define:

UNDEFINED = object()
in library constants.py

Then code above is very clear and I think this is a good and 
sustainable solution.


I know, that in a way ’None’ is already this, but it is generally used 
in 2 ways:

1. A variable/argument is undefined in python space
2. Global indication, that a value is undefined (e.g. equivalence of 
null in json)


What I am proposing is to have 1 more constant in python ‘Undefined’, 
which will indicate ‘Undefined’ in python space and leave None to be 
used in more general sense.


I also found this article while looking for solution just now:
https://levelup.gitconnected.com/python-why-none-is-not-nothing-bb3de55dd471
It seems that people who use typing are also running into ambiguities 
and trying to hack something together.

I avoid typing until its 99% mature so can't comment more on this.

What are your thoughts? Am I onto something or am I missing something? 
Anyone else had similar encounters?









___
Python-ideas mailing list --python-ideas@python.org
To unsubscribe send an email topython-ideas-le...@python.org
https://mail.python.org/mailman3/lists/python-ideas.python.org/
Message archived 
athttps://mail.python.org/archives/list/python-ideas@python.org/message/RGUNJYXWHE7UM2W6FCAMGLJVLXEYHA2Q/
Code of Conduct:http://python.org/psf/codeofconduct/
___
Python-ideas mailing list -- python-ideas@python.org
To unsubscribe send an email to python-ideas-le...@python.org
https://mail.python.org/mailman3/lists/python-ideas.python.org/
Message archived at 
https://mail.python.org/archives/list/python-ideas@python.org/message/2EH4RSB6LPZLV7OOQOOI65DWFSKCZNS2/
Code of Conduct: http://python.org/psf/codeofconduct/

[Python-ideas] Re: Undefined type

[Michael Foord]

There's an implementation of a sentinel object in unittest.mock :

>>> from unittest.mock import sentinel
>>> sentinel.THING
sentinel.THING

On Thu, 8 Jun 2023 at 15:20, David Mertz, Ph.D. 
wrote:

> https://peps.python.org/pep-0661/
>
> On Thu, Jun 8, 2023 at 9:58 AM Dom Grigonis 
> wrote:
>
>> This has been bugging me for a long time. It seems that python doesn’t
>> have a convention for “Undefined” type.
>>
>> When I started using python I naturally used None for it (as advised).
>>
>> However, the more I work with python, the more I find that None is a
>> meaningful type. And it is a type that is convenient to use in a lot of
>> cases, where it is not actually “None”.
>>
>> One big example is:
>> numpy: array[None, :]
>>
>> Another is simple cascading of defaults.
>> E.g.
>>
>> ```python
>> class A:
>> def __init__(self, default=’string'):
>> self.default = default
>>
>> def method(self, default=None):
>> default = default if default is not None else self.default
>> dct[‘key’] = other_dict.get(‘key’, default)
>> return to_json(dct)
>> ```
>>
>> None has a logic in `some_func`, which is important.
>> Now if I want to enforce None passed to to_json I have no way to do it.
>>
>> I know there are many workarounds for this, but there is no `consistent`
>> way of doing this.
>>
>> I have been using Ellipsis in the past, but it doesn’t feel right,
>> because it also has a meaning and I might need to use those functions with
>> e.g. numpy in he future and will have to change it.
>>
>> Now, I came back to this issue again and my latest solution is to just
>> define:
>> UNDEFINED = object()
>> in library constants.py
>>
>> Then code above is very clear and I think this is a good and sustainable
>> solution.
>>
>> I know, that in a way ’None’ is already this, but it is generally used in
>> 2 ways:
>> 1. A variable/argument is undefined in python space
>> 2. Global indication, that a value is undefined (e.g. equivalence of null
>> in json)
>>
>> What I am proposing is to have 1 more constant in python ‘Undefined’,
>> which will indicate ‘Undefined’ in python space and leave None to be used
>> in more general sense.
>>
>> I also found this article while looking for solution just now:
>>
>> https://levelup.gitconnected.com/python-why-none-is-not-nothing-bb3de55dd471
>> It seems that people who use typing are also running into ambiguities and
>> trying to hack something together.
>> I avoid typing until its 99% mature so can't comment more on this.
>>
>> What are your thoughts? Am I onto something or am I missing something?
>> Anyone else had similar encounters?
>>
>>
>>
>>
>>
>>
>>
>>
>> ___
>> Python-ideas mailing list -- python-ideas@python.org
>> To unsubscribe send an email to python-ideas-le...@python.org
>> https://mail.python.org/mailman3/lists/python-ideas.python.org/
>> Message archived at
>> https://mail.python.org/archives/list/python-ideas@python.org/message/RGUNJYXWHE7UM2W6FCAMGLJVLXEYHA2Q/
>> Code of Conduct: http://python.org/psf/codeofconduct/
>>
>
>
> --
> The dead increasingly dominate and strangle both the living and the
> not-yet born.  Vampiric capital and undead corporate persons abuse
> the lives and control the thoughts of homo faber. Ideas, once born,
> become abortifacients against new conceptions.
> ___
> Python-ideas mailing list -- python-ideas@python.org
> To unsubscribe send an email to python-ideas-le...@python.org
> https://mail.python.org/mailman3/lists/python-ideas.python.org/
> Message archived at
> https://mail.python.org/archives/list/python-ideas@python.org/message/YTCPDFDUZTMTXS5LYHKCYJ23WVMPE6CL/
> Code of Conduct: http://python.org/psf/codeofconduct/
>


-- 

Michael Foord

Python Consultant, Contractor and Trainer

https://agileabstractions.com/
___
Python-ideas mailing list -- python-ideas@python.org
To unsubscribe send an email to python-ideas-le...@python.org
https://mail.python.org/mailman3/lists/python-ideas.python.org/
Message archived at 
https://mail.python.org/archives/list/python-ideas@python.org/message/YNKTVZM6GMQAGR5XAJ3SH6HTAWGYG5VT/
Code of Conduct: http://python.org/psf/codeofconduct/

[Python-ideas] Re: Undefined type

[David Mertz, Ph.D.]

https://peps.python.org/pep-0661/

On Thu, Jun 8, 2023 at 9:58 AM Dom Grigonis  wrote:

> This has been bugging me for a long time. It seems that python doesn’t
> have a convention for “Undefined” type.
>
> When I started using python I naturally used None for it (as advised).
>
> However, the more I work with python, the more I find that None is a
> meaningful type. And it is a type that is convenient to use in a lot of
> cases, where it is not actually “None”.
>
> One big example is:
> numpy: array[None, :]
>
> Another is simple cascading of defaults.
> E.g.
>
> ```python
> class A:
> def __init__(self, default=’string'):
> self.default = default
>
> def method(self, default=None):
> default = default if default is not None else self.default
> dct[‘key’] = other_dict.get(‘key’, default)
> return to_json(dct)
> ```
>
> None has a logic in `some_func`, which is important.
> Now if I want to enforce None passed to to_json I have no way to do it.
>
> I know there are many workarounds for this, but there is no `consistent`
> way of doing this.
>
> I have been using Ellipsis in the past, but it doesn’t feel right, because
> it also has a meaning and I might need to use those functions with e.g.
> numpy in he future and will have to change it.
>
> Now, I came back to this issue again and my latest solution is to just
> define:
> UNDEFINED = object()
> in library constants.py
>
> Then code above is very clear and I think this is a good and sustainable
> solution.
>
> I know, that in a way ’None’ is already this, but it is generally used in
> 2 ways:
> 1. A variable/argument is undefined in python space
> 2. Global indication, that a value is undefined (e.g. equivalence of null
> in json)
>
> What I am proposing is to have 1 more constant in python ‘Undefined’,
> which will indicate ‘Undefined’ in python space and leave None to be used
> in more general sense.
>
> I also found this article while looking for solution just now:
>
> https://levelup.gitconnected.com/python-why-none-is-not-nothing-bb3de55dd471
> It seems that people who use typing are also running into ambiguities and
> trying to hack something together.
> I avoid typing until its 99% mature so can't comment more on this.
>
> What are your thoughts? Am I onto something or am I missing something?
> Anyone else had similar encounters?
>
>
>
>
>
>
>
>
> ___
> Python-ideas mailing list -- python-ideas@python.org
> To unsubscribe send an email to python-ideas-le...@python.org
> https://mail.python.org/mailman3/lists/python-ideas.python.org/
> Message archived at
> https://mail.python.org/archives/list/python-ideas@python.org/message/RGUNJYXWHE7UM2W6FCAMGLJVLXEYHA2Q/
> Code of Conduct: http://python.org/psf/codeofconduct/
>


-- 
The dead increasingly dominate and strangle both the living and the
not-yet born.  Vampiric capital and undead corporate persons abuse
the lives and control the thoughts of homo faber. Ideas, once born,
become abortifacients against new conceptions.
___
Python-ideas mailing list -- python-ideas@python.org
To unsubscribe send an email to python-ideas-le...@python.org
https://mail.python.org/mailman3/lists/python-ideas.python.org/
Message archived at 
https://mail.python.org/archives/list/python-ideas@python.org/message/YTCPDFDUZTMTXS5LYHKCYJ23WVMPE6CL/
Code of Conduct: http://python.org/psf/codeofconduct/