Proposal: make Model __unicode__() default to self.name

[cmawe...@gmail.com]

Hi All,

Have you ever quickly set up a model, ran syncdb, and added a few sample objects in the admin to only see a bunch of "MyModel object"s in the changelist? I always forget to add a __unicode__()/__str__() method on my models.

I ran "git grep -1 __unicode__" on some of my django projects and noticed a lot of repeated code. In fact, it seems that in about a _third_ of all my cases, I'm just returning self.name, or returning self.name would have been a good default. I looked at a few 3rd party apps for comparison and found similar results, though not for every app.

IMHO, returning self.name (if the field or property exists) is a sensible default for __unicode__. We can still return "MyModel object" if there's no "name" attribute. You'll still end up adding your own __unicode__ method much of the time, just like you always have.

Yes, it's "magic", but we can document it.

Yes, it's a little more confusing, but we don't have to explain it during the tutorial.

Yes, it's backwards incompatible, but only in rare cases should it be a problem.

Yes, it could look like any Model without a "name" field is "wrong", but it's not.

Yes, "title" is also very popular, but name is better. :)

It has the effect of being a little more friendly in many cases, and can result in more DRY code.

What do your __unicode__/__str__ methods look like? Is this a bad idea?

Thanks,

Collin

[Luke Sneeringer]

What makes “name” special. Why not “label”, “title”, “code”, or any number of other things that people use internally?

It also seems like you could solve this very easily by creating a stub model:

    from django.db import models

    class MyModel(models.Model):

        class Meta:

            abstract = True

        def __unicode__(self):

            return self.name

This has the advantage of your not having to repeat yourself often (and it gives you another place to put your own particular things which you consistently practice), while not becoming opinionated about a naming scheme across the entire framework.

Best Regards,

Luke Sneeringer

P. S. A place where I *would* love to see a better Django default would be in the __repr__ method.

-- 
You received this message because you are subscribed to the Google Groups "Django developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to django-develop...@googlegroups.com.
To post to this group, send email to django-d...@googlegroups.com.
Visit this group at http://groups.google.com/group/django-developers.
For more options, visit https://groups.google.com/groups/opt_out.

[Curtis Maloney]

Luke beat me to it...

We have a couple of abstract models which will fall through a small list of field names, trying to guess a useful default for __str__ ... but it makes me feel dirty every time I see it.

Beyond "blessing" a particular (set of?) field names as default str names... this also adds more magic, and the trend I'm seeing now in the IRC channel is people are believing too much in magic.

Just my 0.0000000152 bitcoins...

--

Curtis

[Aymeric Augustin]

On 11 juil. 2013, at 05:46, cmawe...@gmail.com wrote:

> What do your __unicode__/__str__ methods look like? Is this a bad idea?

In addition to the previous answers, I find it important to have an unambiguous __unicode__ / __str__, and name isn't necessarily unique.

--
Aymeric.

[mjl Martin J. Laubach]

> In addition to the previous answers, I find it important to have an unambiguous __unicode__ / __str__, and name isn't necessarily unique.

  Perhaps to have it a tiny little bit more useful, have it default to "MyModel object (object_pk)" ?

        mjl

[Aymeric Augustin]

2013/7/11 mjl Martin J. Laubach <goo...@emsi.priv.at>

 

Perhaps to have it a tiny little bit more useful, have it default to "MyModel object (object_pk)" ?

I could support that. object_pk will be None when the object isn't saved yet, but that's fine.

--
Aymeric.

[Shai Berger]

So if we're already doing that, take the extra step and make it "MyModel
object (unsaved)" in that case.

[Kapil Garg]

Ticket 27953 is regarding this proposal and the suggestion is about adding "pk" in Model string representation if it exists. 

[Collin Anderson]

Makes sense to me. Maybe still keep the "Transaction object" part, and use None if no pk.

--
You received this message because you are subscribed to the Google Groups "Django developers (Contributions to Django itself)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to django-developers+unsubscribe@googlegroups.com.
To post to this group, send email to django-developers@googlegroups.com.
Visit this group at https://groups.google.com/group/django-developers.
To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/c7254a96-7ee3-4262-a90b-83dcfe8fa3d4%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Kapil Garg]

So what should the final __str__ show: Should it be 'ClassName object pk=Something' and if pk is None then should it be 'ClassName object (unsaved)' or 'ClassName object pk=None' ?

On Sunday, 2 April 2017 23:47:01 UTC+5:30, Collin Anderson wrote:

Makes sense to me. Maybe still keep the "Transaction object" part, and use None if no pk.

On Sun, Apr 2, 2017 at 11:09 AM, Kapil Garg <kapilg...@gmail.com> wrote:

Ticket 27953 is regarding this proposal and the suggestion is about adding "pk" in Model string representation if it exists. 

On Thursday, 11 July 2013 09:16:25 UTC+5:30, Collin Anderson wrote:

Hi All,

Have you ever quickly set up a model, ran syncdb, and added a few sample objects in the admin to only see a bunch of "MyModel object"s in the changelist? I always forget to add a __unicode__()/__str__() method on my models.

I ran "git grep -1 __unicode__" on some of my django projects and noticed a lot of repeated code. In fact, it seems that in about a _third_ of all my cases, I'm just returning self.name, or returning self.name would have been a good default. I looked at a few 3rd party apps for comparison and found similar results, though not for every app.

IMHO, returning self.name (if the field or property exists) is a sensible default for __unicode__. We can still return "MyModel object" if there's no "name" attribute. You'll still end up adding your own __unicode__ method much of the time, just like you always have.

Yes, it's "magic", but we can document it.

Yes, it's a little more confusing, but we don't have to explain it during the tutorial.

Yes, it's backwards incompatible, but only in rare cases should it be a problem.

Yes, it could look like any Model without a "name" field is "wrong", but it's not.

Yes, "title" is also very popular, but name is better. :)

It has the effect of being a little more friendly in many cases, and can result in more DRY code.

What do your __unicode__/__str__ methods look like? Is this a bad idea?

Thanks,

Collin

--
You received this message because you are subscribed to the Google Groups "Django developers (Contributions to Django itself)" group.

To unsubscribe from this group and stop receiving emails from it, send an email to django-develop...@googlegroups.com.
To post to this group, send email to django-d...@googlegroups.com.

[Collin Anderson]

I'd recommend not saying "unsaved". "new" if anything. UUID pk's may default to generating a pk before save, so it might just be best to show the actual current pk value

To unsubscribe from this group and stop receiving emails from it, send an email to django-developers+unsubscribe@googlegroups.com.
To post to this group, send email to django-developers@googlegroups.com.

Visit this group at https://groups.google.com/group/django-developers.

To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/1a84fb1c-522f-4bc4-b214-0831a63f13ae%40googlegroups.com.

[Kapil Garg]

So does this patch seem fine ? 

diff --git a/django/db/models/base.py b/django/db/models/base.py

index 3c1cb9a..f58e12b 100644

--- a/django/db/models/base.py

+++ b/django/db/models/base.py

@@ -504,7 +504,7 @@ class Model(metaclass=ModelBase):

         return '<%s: %s>' % (self.__class__.__name__, u)

 

     def __str__(self):

-        return '%s object' % self.__class__.__name__

+        return '%s object pk=%s' % (self.__class__.__name__, self._get_pk_val())

 

     def __eq__(self, other):

         if not isinstance(other, Model):

[Marco Silva]

I noticed this on the init

def __init__(self, *args, **kwargs):
 # Alias some things as locals to avoid repeat global lookups
 cls = self.__class__

maybe you should change it to self.cls??

Try to submit a PR to the open ticket.

[Kapil Garg]

The opened ticket is about Model.__str__ method. Should i open a new ticket for this change ?
As i see in code, self.__class__ is used in a lot of places but will it effect optimization if we change lookups from self.__class__ to self.cls

Because the methods where class is being used frequently, already store it in local variable and then make references to local variable.

So should it really be changed ?

You received this message because you are subscribed to a topic in the Google Groups "Django developers (Contributions to Django itself)" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/django-developers/7Jqzwg5nt-c/unsubscribe.
To unsubscribe from this group and all its topics, send an email to django-develop...@googlegroups.com.

To post to this group, send email to django-d...@googlegroups.com.
Visit this group at https://groups.google.com/group/django-developers.

To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/a776fdc4-4172-45ca-9a00-06016b24413a%40googlegroups.com.

[Collin Anderson]

Hi All,

Looking at the code snippet, cls is a _local_ variable, not an instance attribute (self.cls). If we changed it to an instance attribute, it would lose most of the speed optimization. Also, self.cls would possibly clash with a field named "cls".

Collin

To unsubscribe from this group and all its topics, send an email to django-developers+unsubscribe@googlegroups.com.
To post to this group, send email to django-developers@googlegroups.com.

Visit this group at https://groups.google.com/group/django-developers.
To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/a776fdc4-4172-45ca-9a00-06016b24413a%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to the Google Groups "Django developers (Contributions to Django itself)" group.

To unsubscribe from this group and stop receiving emails from it, send an email to django-developers+unsubscribe@googlegroups.com.
To post to this group, send email to django-developers@googlegroups.com.

Visit this group at https://groups.google.com/group/django-developers.

To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/CABAYvYp0Xn68hLN0Ne%3DGxWGhTFpMMFPYTJDR6cCStpGwRWCcKA%40mail.gmail.com.

[is_null]

Usually people tell me that type(self) is better than self.__class__, so you might prepare to change that too.

But yeah, it seems like a very reasonable change to me, reading lists of indistinct <Model object> have always required extra steps to figure out what models they are.

Thanks for your effort,

Best
James

[Marco Silva]

I have no idea what is the best way, just say that comment. this is the original PR

https://github.com/django/django/commit/d2a26c1a90e837777dabdf3d67ceec4d2a70fb86

I think you should submit the PR to change the __str__ method, and maybe open a new discussion regarding __repr__

[Marco Silva]

just saw that __repr__ is now under discusion here

https://groups.google.com/forum/#!topic/django-developers/UNKFMg6DO5s

[Collin Anderson]

I just made a pull request.

https://github.com/django/django/pull/8336

(1) is my first choice, pk=1 is my second choice. I'd be fine with either.

To unsubscribe from this group and stop receiving emails from it, send an email to django-developers+unsubscribe@googlegroups.com.
To post to this group, send email to django-developers@googlegroups.com.

Visit this group at https://groups.google.com/group/django-developers.

To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/eb14b116-cdc0-44bd-9f83-86edb5cc7525%40googlegroups.com.

[Alex Corcoles]

Sorry to resurrect this, but I commented on the bug tracker ( https://code.djangoproject.com/ticket/28839#comment:4 ) and was told to take the discussion here.

I want to point out that __str__ is used for "GUI" purposes (such as dropdown texts in ModelForm/ModelAdmin) and putting a default __str__ with a PK doesn't seem to make sense. I think __str__ is a more user/human-friendly thing, which I don't think can be generated automatically.

On the other hand, it does make sense to include the PK in __repr__, so I would suggest backtracking on this change and making an equivalent change in __repr__.

[Collin Anderson]

Hi Alex,

> I think __str__ is a more user/human-friendly thing, which I don't think can be generated automatically.

Yes, having a good __str__ is important, and thats why i think defaulting it to the pk is good, because it's better than nothing. Yes, you'll still want to override __str__ with something better (that likely doesn't include the PK).

So if you're overriding __str__ anyway, why not also override __repr__ to include the PK?

Or what are you actually proposing, something like this?

https://github.com/django/django/compare/master...collinanderson:patch-12

Thanks,

Collin

To unsubscribe from this group and stop receiving emails from it, send an email to django-developers+unsubscribe@googlegroups.com.
To post to this group, send email to django-developers@googlegroups.com.

Visit this group at https://groups.google.com/group/django-developers.

To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/a37f50c0-c58f-45ec-aa3c-dcc117fcece1%40googlegroups.com.

[Alex Corcoles]

Hi Collin,

On Tuesday, December 12, 2017 at 3:48:38 PM UTC+1, Collin Anderson wrote:

Yes, having a good __str__ is important, and thats why i think defaulting it to the pk is good, because it's better than nothing. Yes, you'll still want to override __str__ with something better (that likely doesn't include the PK).

I slightly disagree on this specific point, but I don't think my squabble is worth wasting anyone's time :)

So if you're overriding __str__ anyway, why not also override __repr__ to include the PK?

Or what are you actually proposing, something like this?

https://github.com/django/django/compare/master...collinanderson:patch-12

 
Yeah, I just want something like your PR. I want to override __str__ most of the time, so having it include the PK doesn't do much for me- neither good or bad, but having the default __repr__ include the PK saves me from overriding it most of the time.

I could nitpick on my ideal implementation of both methods, but I think it comes down to personal taste so I just wanted to propose having PK in the default __repr__ implementation as I think that will benefit people. The rest doesn't matter that much to me.

Thank you,

Álex