如何解决 Django 中缺乏跨数据库外键支持的问题

Question

我知道 Django 不支持外国跨多个数据库的键（最初是 Django 1.3 文档）

但我正在寻找解决方法。

什么不起作用

我有两个模型，每个模型都在一个单独的数据库上。

routers.py：fruit_app/models.py 中的

class NewsRouter(object):
    def db_for_read(self, model, **hints):
        if model._meta.app_label == 'news_app':
            return 'news_db'
        return None

    def db_for_write(self, model, **hints):
        if model._meta.app_label == 'news_app':
            return 'news_db'
        return None

    def allow_relation(self, obj1, obj2, **hints):
        if obj1._meta.app_label == 'news_app' or obj2._meta.app_label == 'news_app':
            return True
        return None

    def allow_syncdb(self, db, model):
        if db == 'news_db':
            return model._meta.app_label == 'news_app'
        elif model._meta.app_label == 'news_app':
            return False
        return None

模型 1：

from django.db import models

class Fruit(models.Model):
    name = models.CharField(max_length=20)

news_app/models.py 中的模型 2：

from django.db import models

class Article(models.Model):
    fruit = models.ForeignKey('fruit_app.Fruit')
    intro = models.TextField()

尝试在管理中添加“文章”会出现以下错误，因为它正在寻找 Fruit在错误的数据库上建立模型 ('news_db')：

DatabaseError at /admin/news_app/article/add/

(1146, "Table 'fkad_news.fruit_app_fruit' doesn't exist")

方法 1：子类 IntegerField

我创建了一个自定义字段foreignKeyAcrossDb，它是 IntegerField 的子类。代码位于 github 上：https://github。 com/saltycrane/django-foreign-key-across-db-testproject/tree/integerfield_subclass

fields.py：

from django.db import models


class ForeignKeyAcrossDb(models.IntegerField):
    '''
    Exists because foreign keys do not work across databases
    '''
    def __init__(self, model_on_other_db, **kwargs):
        self.model_on_other_db = model_on_other_db
        super(ForeignKeyAcrossDb, self).__init__(**kwargs)

    def to_python(self, value):
        # TODO: this db lookup is duplicated in get_prep_lookup()
        if isinstance(value, self.model_on_other_db):
            return value
        else:
            return self.model_on_other_db._default_manager.get(pk=value)

    def get_prep_value(self, value):
        if isinstance(value, self.model_on_other_db):
            value = value.pk
        return super(ForeignKeyAcrossDb, self).get_prep_value(value)

    def get_prep_lookup(self, lookup_type, value):
        # TODO: this db lookup is duplicated in to_python()
        if not isinstance(value, self.model_on_other_db):
            value = self.model_on_other_db._default_manager.get(pk=value)

        return super(ForeignKeyAcrossDb, self).get_prep_lookup(lookup_type, value)

我将我的文章模型更改为：

class Article(models.Model):
    fruit = ForeignKeyAcrossDb(Fruit)
    intro = models.TextField()

问题是，有时当我访问 Article.fruit 时，它是一个整数，有时是 Fruit 对象。我希望它始终是一个 Fruit 对象。我需要做什么才能使访问 Article.fruit 始终返回 Fruit 对象？

作为我的解决方法的解决方法，我添加了一个 fruit_obj 属性，但如果可能的话，我想消除此属性：

class Article(models.Model):
    fruit = ForeignKeyAcrossDb(Fruit)
    intro = models.TextField()

    # TODO: shouldn't need fruit_obj if ForeignKeyAcrossDb field worked properly
    @property
    def fruit_obj(self):
        if not hasattr(self, '_fruit_obj'):
            # TODO: why is it sometimes an int and sometimes a Fruit object?
            if isinstance(self.fruit, int) or isinstance(self.fruit, long):
                print 'self.fruit IS a number'
                self._fruit_obj = Fruit.objects.get(pk=self.fruit)
            else:
                print 'self.fruit IS NOT a number'
                self._fruit_obj = self.fruit
        return self._fruit_obj

    def fruit_name(self):
        return self.fruit_obj.name

方法 2：子类ForeignKey 字段

作为第二次尝试，我尝试对ForeignKey 字段进行子类化。我修改了 ReverseSingleRelatedObjectDescriptor 以使用 Fruit 模型管理器上的 forced_using 指定的数据库。我还删除了 ForeignKey 子类上的 validate() 方法。该方法没有与方法1相同的问题。github上的代码位于： https://github.com/saltycrane/django-foreign-key-across-db-testproject/tree/foreignkey_subclass

fields.py：fruit_app/models.py：news_app/models.py

from django.db import models
from django.db import router
from django.db.models.query import QuerySet


class ReverseSingleRelatedObjectDescriptor(object):
    # This class provides the functionality that makes the related-object
    # managers available as attributes on a model class, for fields that have
    # a single "remote" value, on the class that defines the related field.
    # In the example "choice.poll", the poll attribute is a
    # ReverseSingleRelatedObjectDescriptor instance.
    def __init__(self, field_with_rel):
        self.field = field_with_rel

    def __get__(self, instance, instance_type=None):
        if instance is None:
            return self

        cache_name = self.field.get_cache_name()
        try:
            return getattr(instance, cache_name)
        except AttributeError:
            val = getattr(instance, self.field.attname)
            if val is None:
                # If NULL is an allowed value, return it.
                if self.field.null:
                    return None
                raise self.field.rel.to.DoesNotExist
            other_field = self.field.rel.get_related_field()
            if other_field.rel:
                params = {'%s__pk' % self.field.rel.field_name: val}
            else:
                params = {'%s__exact' % self.field.rel.field_name: val}

            # If the related manager indicates that it should be used for
            # related fields, respect that.
            rel_mgr = self.field.rel.to._default_manager
            db = router.db_for_read(self.field.rel.to, instance=instance)
            if getattr(rel_mgr, 'forced_using', False):
                db = rel_mgr.forced_using
                rel_obj = rel_mgr.using(db).get(**params)
            elif getattr(rel_mgr, 'use_for_related_fields', False):
                rel_obj = rel_mgr.using(db).get(**params)
            else:
                rel_obj = QuerySet(self.field.rel.to).using(db).get(**params)
            setattr(instance, cache_name, rel_obj)
            return rel_obj

    def __set__(self, instance, value):
        raise NotImplementedError()

class ForeignKeyAcrossDb(models.ForeignKey):

    def contribute_to_class(self, cls, name):
        models.ForeignKey.contribute_to_class(self, cls, name)
        setattr(cls, self.name, ReverseSingleRelatedObjectDescriptor(self))
        if isinstance(self.rel.to, basestring):
            target = self.rel.to
        else:
            target = self.rel.to._meta.db_table
        cls._meta.duplicate_targets[self.column] = (target, "o2m")

    def validate(self, value, model_instance):
        pass

：

from django.db import models


class FruitManager(models.Manager):
    forced_using = 'default'


class Fruit(models.Model):
    name = models.CharField(max_length=20)

    objects = FruitManager()

“

from django.db import models

from foreign_key_across_db_testproject.fields import ForeignKeyAcrossDb
from foreign_key_across_db_testproject.fruit_app.models import Fruit


class Article(models.Model):
    fruit = ForeignKeyAcrossDb(Fruit)
    intro = models.TextField()

    def fruit_name(self):
        return self.fruit.name

方法 2a：为 Fruit_app 添加路由器

此解决方案为 fruit_app 使用额外的路由器。此解决方案不需要对方法 2 中所需的 ForeignKey 进行修改。在 django.db.utils.ConnectionRouter，我们发现即使我们期望 < code>fruit_app 默认位于 'default' 数据库中，将 instance 提示传递给 db_for_read 进行外键查找它位于'news_db' 数据库中。我们添加了第二个路由器，以确保始终从 'default' 数据库读取 fruit_app 模型。 ForeignKey 子类仅用于“修复”ForeignKey.validate() 方法。 （如果 Django 想支持跨数据库的外键，我会说这是 Django 的 bug。） 代码位于 github 上： https://github.com/saltycrane/ django-foreign-key-across-db-testproject

:

class NewsRouter(object):
    def db_for_read(self, model, **hints):
        if model._meta.app_label == 'news_app':
            return 'news_db'
        return None

    def db_for_write(self, model, **hints):
        if model._meta.app_label == 'news_app':
            return 'news_db'
        return None

    def allow_relation(self, obj1, obj2, **hints):
        if obj1._meta.app_label == 'news_app' or obj2._meta.app_label == 'news_app':
            return True
        return None

    def allow_syncdb(self, db, model):
        if db == 'news_db':
            return model._meta.app_label == 'news_app'
        elif model._meta.app_label == 'news_app':
            return False
        return None


class FruitRouter(object):
    def db_for_read(self, model, **hints):
        if model._meta.app_label == 'fruit_app':
            return 'default'
        return None

    def db_for_write(self, model, **hints):
        if model._meta.app_label == 'fruit_app':
            return 'default'
        return None

    def allow_relation(self, obj1, obj2, **hints):
        if obj1._meta.app_label == 'fruit_app' or obj2._meta.app_label == 'fruit_app':
            return True
        return None

    def allow_syncdb(self, db, model):
        if db == 'default':
            return model._meta.app_label == 'fruit_app'
        elif model._meta.app_label == 'fruit_app':
            return False
        return None

routers.py:fruit_app/models.py:news_app/models.py:fields.py

from django.db import models


class Fruit(models.Model):
    name = models.CharField(max_length=20)

django -

from django.db import models

from foreign_key_across_db_testproject.fields import ForeignKeyAcrossDb
from foreign_key_across_db_testproject.fruit_app.models import Fruit


class Article(models.Model):
    fruit = ForeignKeyAcrossDb(Fruit)
    intro = models.TextField()

    def fruit_name(self):
        return self.fruit.name

附加

from django.core import exceptions
from django.db import models
from django.db import router


class ForeignKeyAcrossDb(models.ForeignKey):

    def validate(self, value, model_instance):
        if self.rel.parent_link:
            return
        models.Field.validate(self, value, model_instance)
        if value is None:
            return

        using = router.db_for_read(self.rel.to, instance=model_instance)  # is this more correct than Django's 1.2.5 version?
        qs = self.rel.to._default_manager.using(using).filter(
                **{self.rel.field_name: value}
             )
        qs = qs.complex_filter(self.rel.limit_choices_to)
        if not qs.exists():
            raise exceptions.ValidationError(self.error_messages['invalid'] % {
                'model': self.rel.to._meta.verbose_name, 'pk': value})

信息

• users 列表上的线程有很多信息：http://groups.google.com/group/django -users/browse_thread/thread/74bcd1afdeb2f0/0fdfce061124b915
• 多数据库文档的修订历史记录：http://code.djangoproject.com/log/django/trunk/docs/topics/db/multi-db.txt?verbose=on

更新

在对路由器进行更多调整后，我们实现了最后一种方法。整个实施过程非常痛苦，这让我们认为我们一定做错了。 TODO 列表中正在为此编写单元测试。

Answer 1

您可以在数据库中创建一个包含跨数据库查询的视图，然后在单独的文件中定义该视图的模型以保持syncdb 正常工作。

快乐的编程。 :)

Answer 2

我知道 Djano-nosql 支持键，尽管有一些来自 http://www.allbuttonspressed 的魔法.com/projects/django-dbindexer。也许其中一些可以有所帮助。

根据描述：

“您只需告诉 dbindexer 哪些模型和字段应该支持这些查询，它就会为您维护所需的索引。”

-凯瑞

Answer 3

至于ForeignKeyAcrossDb部分，你不能对__init__中的类进行一些调整吗？检查适当的字段是否为Integer，如果不是，请从数据库加载它，或执行任何其他需要的操作。 Python __class__es 可以在运行时更改，没有太大问题。

Answer 4

在我伤透了几天之后，我设法在同一个银行得到了我的外键！

可以更改表格以在不同的银行寻求外键！

首先，添加一个RECHARGE的FIELDS，都直接（破解）我的表单，在函数____init____

app.form.py

# -*- coding: utf-8 -*-
from django import forms
import datetime
from app_ti_helpdesk import models as mdp

#classe para formulario de Novo HelpDesk
class FormNewHelpDesk(forms.ModelForm):
    class Meta:
        model = mdp.TblHelpDesk
        fields = (
        "problema_alegado",
        "cod_direcionacao",
        "data_prevista",
        "hora_prevista",
        "atendimento_relacionado_a",
        "status",
        "cod_usuario",
        )

    def __init__(self, *args, **kwargs):
        #-------------------------------------
        #  using remove of kwargs
        #-------------------------------------
        db = kwargs.pop("using", None)

        # CASE use Unique Keys
        self.Meta.model.db = db

        super(FormNewHelpDesk, self).__init__(*args,**kwargs)

        #-------------------------------------
        #   recreates the fields manually
        from copy import deepcopy
        self.fields.update(deepcopy( forms.fields_for_model( self.Meta.model, self.Meta.fields, using=db ) ))
        #
        #-------------------------------------

        #### follows the standard template customization, if necessary

        self.fields['problema_alegado'].widget.attrs['rows'] = 3
        self.fields['problema_alegado'].widget.attrs['cols'] = 22
        self.fields['problema_alegado'].required = True
        self.fields['problema_alegado'].error_messages={'required': 'Necessário informar o motivo da solicitação de ajuda!'}


        self.fields['data_prevista'].widget.attrs['class'] = 'calendario'
        self.fields['data_prevista'].initial = (datetime.timedelta(4)+datetime.datetime.now().date()).strftime("%Y-%m-%d")

        self.fields['hora_prevista'].widget.attrs['class'] = 'hora'
        self.fields['hora_prevista'].initial =datetime.datetime.now().time().strftime("%H:%M")

        self.fields['status'].initial = '0'                 #aberto
        self.fields['status'].widget.attrs['disabled'] = True

        self.fields['atendimento_relacionado_a'].initial = '07'

        self.fields['cod_direcionacao'].required = True
        self.fields['cod_direcionacao'].label = "Direcionado a"
        self.fields['cod_direcionacao'].initial = '2'
        self.fields['cod_direcionacao'].error_messages={'required': 'Necessário informar para quem é direcionado a ajuda!'}

        self.fields['cod_usuario'].widget = forms.HiddenInput()

调用表单，

中从视图app.view.py

form = forms.FormNewHelpDesk(request.POST or None, using=banco)

现在，在源代码DJANGO中更改

ForeignKey类型的字段， ManyToManyField 和 OneToOneField 可以使用“using”，因此添加了一个 IF ...

django.forms.models.py

# line - 133: add using=None
def fields_for_model(model, fields=None, exclude=None, widgets=None, formfield_callback=None, using=None):

# line - 159

if formfield_callback is None:
    #----------------------------------------------------
    from django.db.models.fields.related import (ForeignKey, ManyToManyField, OneToOneField)
    if type(f) in (ForeignKey, ManyToManyField, OneToOneField):
        kwargs['using'] = using

    formfield = f.formfield(**kwargs)
    #----------------------------------------------------
elif not callable(formfield_callback):
    raise TypeError('formfield_callback must be a function or callable')
else:
    formfield = formfield_callback(f, **kwargs)

ALTER FOLLOW FILE

django.db.models.base.py

alter

# line 717
qs = model_class._default_manager.filter(**lookup_kwargs)

for

# line 717
qs = model_class._default_manager.using(getattr(self, 'db', None)).filter(**lookup_kwargs)

Ready :D

Answer 5

外键字段意味着您可以
- 通过加入iefruit__name查询关系
- 检查引用完整性
- 确保删除时的引用完整性
- 管理原始ID查找功能
-（更多...）

第一个用例总是有问题的。
代码库中可能还有一些其他外键特殊情况也不起作用。

我运行一个相当大的 django 站点，我们目前使用一个普通的整数字段。
现在我认为子类化整数字段并将 id 添加到对象转换将是最简单的（在 1.2 中需要修补 django 的一些位，希望现在有所改进）
让您知道我们找到了什么解决方案。

Answer 6

遇到了类似的问题，需要跨多个 (5) 个数据库引用（大部分）静态数据。对 ReversedSingleRelatedObjectDescriptor 进行了轻微更新，以允许设置相关模型。它没有实现反向关系atm。

class ReverseSingleRelatedObjectDescriptor(object):
"""
This class provides the functionality that makes the related-object managers available as attributes on a model
class, for fields that have a single "remote" value, on the class that defines the related field. Used with
LinkedField.
"""
def __init__(self, field_with_rel):
    self.field = field_with_rel
    self.cache_name = self.field.get_cache_name()

def __get__(self, instance, instance_type=None):
    if instance is None:
        return self

    try:
        return getattr(instance, self.cache_name)
    except AttributeError:
        val = getattr(instance, self.field.attname)
        if val is None:
            # If NULL is an allowed value, return it
            if self.field.null:
                return None
            raise self.field.rel.to.DoesNotExist
        other_field = self.field.rel.get_related_field()
        if other_field.rel:
            params = {'%s__pk' % self.field.rel.field_name: val}
        else:
            params = {'%s__exact' % self.field.rel.field_name: val}

        # If the related manager indicates that it should be used for related fields, respect that.
        rel_mgr = self.field.rel.to._default_manager
        db = router.db_for_read(self.field.rel.to, instance=instance)
        if getattr(rel_mgr, 'forced_using', False):
            db = rel_mgr.forced_using
            rel_obj = rel_mgr.using(db).get(**params)
        elif getattr(rel_mgr, 'use_for_related_fields', False):
            rel_obj = rel_mgr.using(db).get(**params)
        else:
            rel_obj = QuerySet(self.field.rel.to).using(db).get(**params)
        setattr(instance, self.cache_name, rel_obj)
        return rel_obj

def __set__(self, instance, value):
    if instance is None:
        raise AttributeError("%s must be accessed via instance" % self.field.name)

    # If null=True, we can assign null here, but otherwise the value needs to be an instance of the related class.
    if value is None and self.field.null is False:
        raise ValueError('Cannot assign None: "%s.%s" does not allow null values.' %
                         (instance._meta.object_name, self.field.names))
    elif value is not None and not isinstance(value, self.field.rel.to):
        raise ValueError('Cannot assign "%r": "%s.%s" must be a "%s" instance.' %
                         (value, instance._meta.object_name, self.field.name, self.field.rel.to._meta.object_name))
    elif value is not None:
        # Only check the instance state db, LinkedField implies that the value is on a different database
        if instance._state.db is None:
            instance._state.db = router.db_for_write(instance.__class__, instance=value)

    # Is not used by OneToOneField, no extra measures to take here

    # Set the value of the related field
    try:
        val = getattr(value, self.field.rel.get_related_field().attname)
    except AttributeError:
        val = None
    setattr(instance, self.field.attname, val)

    # Since we already know what the related object is, seed the related object caches now, too. This avoids another
    # db hit if you get the object you just set
    setattr(instance, self.cache_name, value)
    if value is not None and not self.field.rel.multiple:
        setattr(value, self.field.related.get_cache_name(), instance)

和

class LinkedField(models.ForeignKey):
"""
Field class used to link models across databases. Does not ensure referrential integraty like ForeignKey
"""
def _description(self):
    return "Linked Field (type determined by related field)"

def contribute_to_class(self, cls, name):
    models.ForeignKey.contribute_to_class(self, cls, name)
    setattr(cls, self.name, ReverseSingleRelatedObjectDescriptor(self))
    if isinstance(self.rel.to, basestring):
        target = self.rel.to
    else:
        target = self.rel.to._meta.db_table
    cls._meta.duplicate_targets[self.column] = (target, "o2m")

def validate(self, value, model_instance):
    pass

Answer 7

该解决方案最初是为一个具有迁移功能的托管数据库以及一个或多个具有在数据库级别连接到同一数据库的模型 Meta managed=False 的旧数据库而编写的。如果db_table选项包含数据库名称加上表名称正确引用'`'（MySQL）或' " '（其他db），例如db_table = '"DB2"."table_b"'，则Django不再引用它。查询由Django编译正确的 ORM，即使是JOIN：

class TableB(models.Model):
    ....
    class Meta:    
        db_table = '`DB2`.`table_b`'    # for MySQL
        # db_table = '"DB2"."table_b"'  # for all other backends
        managed = False

查询集：

>>> qs = TableB.objects.all()
>>> str(qs.query)
'SELECT "DB2"."table_b"."id" FROM DB2"."table_b"'

Django 中的所有数据库后端都支持它

（看来我在 重复新问题，我的答案继续。）

Answer 8

受到@Frans 评论的启发。我的解决方法是在业务层执行此操作。在给出这个问题的例子中。我会将结果设置为 Article 上的 IntegerField，作为“不在数据层中进行完整性检查”。

class Fruit(models.Model):
    name = models.CharField()

class Article(models.Model):
    fruit = models.IntegerField()
    intro = models.TextField()

然后尊重应用程序代码（业务层）中的引用关系。以 Django admin 为例，为了在文章的添加页面中显示水果作为选项，您需要手动填充水果的选项列表。

# admin.py in App article
class ArticleAdmin(admin.ModelAdmin):
    class ArticleForm(forms.ModelForm):
        fields = ['fruit', 'intro']

        # populate choices for fruit
        choices = [(obj.id, obj.name) for obj in Fruit.objects.all()]
        widgets = {
            'fruit': forms.Select(choices=choices)}

    form = ArticleForm
    list_diaplay = ['fruit', 'intro']

当然，您可能需要处理表单字段验证（完整性检查）。

Answer 9

我有一个 django v1.10 的新解决方案。有两个部分。它与 django.admin 和 django.rest-framework 一起使用。

1. 继承ForeignKey类并创建ForeignKeyAcrossDb，并基于此票 和此 帖子。

class ForeignKeyAcrossDb(models.ForeignKey):
        def validate(self, value, model_instance):
            if self.remote_field.parent_link:
                return
            super(models.ForeignKey, self).validate(value, model_instance)
            if value is None:
                return
            using = router.db_for_read(self.remote_field.model, instance=model_instance)
            qs = self.remote_field.model._default_manager.using(using).filter(
                **{self.remote_field.field_name: value}
            )
            qs = qs.complex_filter(self.get_limit_choices_to())
            if not qs.exists():
                raise exceptions.ValidationError(
                    self.error_messages['invalid'],
                    code='invalid',
                    params={
                        'model': self.remote_field.model._meta.verbose_name, 'pk': value,
                        'field': self.remote_field.field_name, 'value': value,
                    },  # 'pk' is included for backwards compatibility
                )

2. 在字段声明中，使用db_constraint=False，例如

album=ForeignKeyAcrossDb(Singer, db_constraint=False, on_delete=models.DO_NOTHING)