You are trying to add a non-nullable field 'new_field' to userprofile without a default

If “website” can be empty than new_field should also be set to be empty.

Now if you want to add logic on save where if new_field is empty to grab the value from “website” all you need to do is override the save function for your Model like this:

class UserProfile(models.Model):
    user = models.OneToOneField(User)
    website = models.URLField(blank=True, default='DEFAULT VALUE')
    new_field = models.CharField(max_length=140, blank=True, default='DEFAULT VALUE')

    def save(self, *args, **kwargs):
        if not self.new_field:
            # Setting the value of new_field with website's value
            self.new_field = self.website

        # Saving the object with the default save() function
        super(UserProfile, self).save(*args, **kwargs)

One option is to declare a default value for ‘new_field’:

new_field = models.CharField(max_length=140, default='DEFAULT VALUE')

another option is to declare ‘new_field’ as a nullable field:

new_field = models.CharField(max_length=140, null=True)

If you decide to accept ‘new_field’ as a nullable field you may want to accept ‘no input’ as valid input for ‘new_field’. Then you have to add the blank=True statement as well:

new_field = models.CharField(max_length=140, blank=True, null=True)

Even with null=True and/or blank=True you can add a default value if necessary:

new_field = models.CharField(max_length=140, default='DEFAULT VALUE', blank=True, null=True)

In new_file add the boolean property null.

new_field = models.CharField(max_length=140, null=True)

after you run a ./manage.py syncdb for refresh the DB.
and finally you run ./manage.py makemigrations
and ./manage.py migrate

If you are in early development cycle and don’t care about your current database data you can just remove it and then migrate. But first you need to clean migrations dir and remove its rows from table (django_migrations)

rm  your_app/migrations/* 

Note: Don’t delete _ _ init _ _ .py in the migrations folder.

rm db.sqlite3
python manage.py makemigrations
python manage.py migrate

If you previously created a superuser for Django’s Admin web app then you might need to create the user again.

Do you already have database entries in the table UserProfile? If so, when you add new columns the DB doesn’t know what to set it to because it can’t be NULL. Therefore it asks you what you want to set those fields in the column new_fields to. I had to delete all the rows from this table to solve the problem.

(I know this was answered some time ago, but I just ran into this problem and this was my solution. Hopefully it will help anyone new that sees this)

You can use method from Django Doc from this page https://docs.djangoproject.com/en/1.8/ref/models/fields/#default

Create default and use it

def contact_default():
   return {"email": "[email protected]"}

contact_info = JSONField("ContactInfo", default=contact_default)

If you are early into the development cycle you can try this –

Remove/comment that model and all its usages. Apply migrations. That would delete that model and then add the model again, run migrations and you have a clean model with the new field added.

What Django actually says is:

Userprofile table has data in it and there might be new_field values
which are null, but I do not know, so are you sure you want to mark
property as non nullable, because if you do you might get an error if
there are values with NULL

If you are sure that none of values in the userprofile table are NULL – fell free and ignore the warning.

The best practice in such cases would be to create a RunPython migration to handle empty values as it states in option 2

2) Ignore for now, and let me handle existing rows with NULL myself (e.g. because you added a RunPython or RunSQL operation to handle NULL values in a previous data migration)

In RunPython migration you have to find all UserProfile instances with empty new_field value and put a correct value there (or a default value as Django asks you to set in the model).
You will get something like this:

# please keep in mind that new_value can be an empty string. You decide whether it is a correct value.
for profile in UserProfile.objects.filter(new_value__isnull=True).iterator():
    profile.new_value = calculate_value(profile)
    profile.save() # better to use batch save

Have fun!

I honestly fount the best way to get around this was to just create another model with all the fields that you require and named slightly different. Run migrations. Delete unused model and run migrations again. Voila.

If the SSH it gives you 2 options, choose number 1, and put “None”. Just that…for the moment.

You can’t add reference to table that have already data inside.
Change:

user = models.OneToOneField(User)

to:

user = models.OneToOneField(User, default = "")

do:

python manage.py makemigrations
python manage.py migrate

change again:

user = models.OneToOneField(User)

do migration again:

python manage.py makemigrations
python manage.py migrate

If you are fine with truncating the table of the model in question, you can specify a one-off default value of None in the prompt. The migration will have superfluous default=None while your code has no default. It can be applied just fine because there’s no data in the table anymore which would require a default.

In models.py

class UserProfile(models.Model):
user = models.OneToOneField(User)
website = models.URLField(blank=True)
new_field = models.CharField(max_length=140, default="some_value")

You need to add some values as default.

I was early in my development cycle, so this may not work for everyone (but I don’t see why it wouldn’t).

I added blank=True, null=True to the columns where I was getting the error. Then I ran the python manage.py makemigrations command.

Immediately after running this command (and before running python manage.py migrate), I removed the blank=True, null=True from all the columns. Then I ran python manage.py makemigrations again. I was given an option to just change the columns myself, which I selected.

Then I ran python manage.py migrate and everything worked well!

In case anyone is setting a ForeignKey, you can just allow nullable fields without setting a default:

new_field = models.ForeignKey(model, null=True)

If you already have data stored within the database, you can also set a default value:

new_field = models.ForeignKey(model, default=<existing model id here>)

Basically I solved this Issue putting null=True in added model.
For example:

I was trying to add phone_number in the existing database module.

phone_number = models.CharField(max_length=10)

I previously got this error:

You are trying to add a non-nullable field ‘phone_number’ to customer
without a default; we can’t do that (the database needs something to
populate existing rows). Please select a fix:

1. Provide a one-off
   default now (will be set on all existing rows with a null value for
   this column)
2. Quit, and let me add a default in models.py

After putting

phone_number = models.CharField(max_length=10, null=True)

python manage.py makemigrations
Migrations for 'collegekhajagharapp':
  collegekhajagharappmigrations002_customer_phone_number.py
    - Add field phone_number to customer

If your table is new and you dont care about the data follow below steps:

1. Delete FROM public.django_migrations where app=<your_app_name>
2. Delete all migrations file from your project
3. DROP table <your_table_name>
4. python manage.py makemigrations
5. python manage.py migrate

If you care about data, then add a default to your field

new_field = models.CharField(max_length=140, default='your_default_value')

Here is a workaround without having to make compromises such as dropping all existing data/migrations (yikes), requiring two separate migrations, or setting an unwanted default. Take these steps:

1. Add the new field to your model with null=False
2. python manage.py makemigrations <app_name> --name <migration_name>
3. Change the new field to null=True
4. python manage.py makemigrations <app_name>. In this step, you’ll see an option you may never have seen before! You want option 2: "Ignore for now, and let me handle existing rows with NULL myself (e.g. because you added a RunPython or RunSQL operation to handle NULL values in a previous data migration)"
5. Move the migrations.AlterField operation from the second migration into the first one, and add a manual migration to take care of the NULL values (if you don’t take this step you’ll get integrity errors like before), like so:

def initial_display_names(apps, schema):
    Player = apps.get_model('my_app', 'Player')
    Player.objects.all().update(display_name='Cool Name')

def reversal(*args):
    """Reversal is NOOP since display_name is simply dropped during reverse"""


class Migration(migrations.Migration):

    dependencies = [
        ('my_app', '0049_foo'),
    ]

    operations = [
        migrations.AddField(
            model_name='player',
            name='display_name',
            field=models.CharField(help_text='name as shown in the UI', max_length=256, null=True),
        ),
        migrations.RunPython(initial_display_names, reversal),
        migrations.AlterField(
            model_name='player',
            name='display_name',
            field=models.CharField(help_text='name as shown in the UI', max_length=256),
        ),
    ]

6. Remove the second migration you made.
7. python manage.py migrate <app_name>

Delete Model => Makemigrations Models => Migrate Models => Return Back the models => Makemigrations Models => Migrate Models