多対多の関係を定義するには、ManyToManyField を使用します。
この例では、一つの Article は複数の Publication オブジェクトに掲載することができ、一つの Publication は複数の Article オブジェクトを持ちます:
from django.db import models\n\n\nclass Publication(models.Model):\n title = models.CharField(max_length=30)\n\n class Meta:\n ordering = ["title"]\n\n def __str__(self):\n return self.title\n\n\nclass Article(models.Model):\n headline = models.CharField(max_length=100)\n publications = models.ManyToManyField(Publication)\n\n class Meta:\n ordering = ["headline"]\n\n def __str__(self):\n return self.headline\n以下は、Python API の機能を使って実行できる操作の例です。
\nいくつかの 出版物 を作成する:
>>> p1 = Publication(title="The Python Journal")\n>>> p1.save()\n>>> p2 = Publication(title="Science News")\n>>> p2.save()\n>>> p3 = Publication(title="Science Weekly")\n>>> p3.save()\n一つの Article を作成する:
>>> a1 = Article(headline="Django lets you build web apps easily")\n保存されるまでは Publication と関連付けることはできません:
>>> a1.publications.add(p1)\nTraceback (most recent call last):\n...\nValueError: "<Article: Django lets you build web apps easily>" needs to have a value for field "id" before this many-to-many relationship can be used.\n保存しましょう！
\n>>> a1.save()\nArticle と Publication を関連付ける:
>>> a1.publications.add(p1)\n別の Article を作成し、それを Publications に表示するようにセットする:
>>> a2 = Article(headline="NASA uses Python")\n>>> a2.save()\n>>> a2.publications.add(p1, p2)\n>>> a2.publications.add(p3)\n2回の追加はOKで、関係が重複することはありません:
\n>>> a2.publications.add(p3)\n間違った型のオブジェクトを追加すると TypeError が発生します:
>>> a2.publications.add(a1)\nTraceback (most recent call last):\n...\nTypeError: 'Publication' instance expected\nArticle に Publication をワンステップで作成・追加するには create() を使います:
>>> new_publication = a2.publications.create(title="Highlights for Children")\nArticle オブジェクトはリレーション先の Publication オブジェクトにアクセスできます:
>>> a1.publications.all()\n<QuerySet [<Publication: The Python Journal>]>\n>>> a2.publications.all()\n<QuerySet [<Publication: Highlights for Children>, <Publication: Science News>, <Publication: Science Weekly>, <Publication: The Python Journal>]>\nPublication オブジェクトはリレーション先の Article オブジェクトにアクセスできます:
>>> p2.article_set.all()\n<QuerySet [<Article: NASA uses Python>]>\n>>> p1.article_set.all()\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA uses Python>]>\n>>> Publication.objects.get(id=4).article_set.all()\n<QuerySet [<Article: NASA uses Python>]>\n多対多の関係は リレーションを横断するルックアップ を使ってクエリできます:
\n>>> Article.objects.filter(publications__id=1)\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA uses Python>]>\n>>> Article.objects.filter(publications__pk=1)\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA uses Python>]>\n>>> Article.objects.filter(publications=1)\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA uses Python>]>\n>>> Article.objects.filter(publications=p1)\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA uses Python>]>\n\n>>> Article.objects.filter(publications__title__startswith="Science")\n<QuerySet [<Article: NASA uses Python>, <Article: NASA uses Python>]>\n\n>>> Article.objects.filter(publications__title__startswith="Science").distinct()\n<QuerySet [<Article: NASA uses Python>]>\ncount() 関数は distinct() も尊重します:
>>> Article.objects.filter(publications__title__startswith="Science").count()\n2\n\n>>> Article.objects.filter(publications__title__startswith="Science").distinct().count()\n1\n\n>>> Article.objects.filter(publications__in=[1, 2]).distinct()\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA uses Python>]>\n>>> Article.objects.filter(publications__in=[p1, p2]).distinct()\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA uses Python>]>\n逆の多対多クエリもサポートされています (つまり、 ManyToManyField を持たないテーブルから開始できます):
>>> Publication.objects.filter(id=1)\n<QuerySet [<Publication: The Python Journal>]>\n>>> Publication.objects.filter(pk=1)\n<QuerySet [<Publication: The Python Journal>]>\n\n>>> Publication.objects.filter(article__headline__startswith="NASA")\n<QuerySet [<Publication: Highlights for Children>, <Publication: Science News>, <Publication: Science Weekly>, <Publication: The Python Journal>]>\n\n>>> Publication.objects.filter(article__id=1)\n<QuerySet [<Publication: The Python Journal>]>\n>>> Publication.objects.filter(article__pk=1)\n<QuerySet [<Publication: The Python Journal>]>\n>>> Publication.objects.filter(article=1)\n<QuerySet [<Publication: The Python Journal>]>\n>>> Publication.objects.filter(article=a1)\n<QuerySet [<Publication: The Python Journal>]>\n\n>>> Publication.objects.filter(article__in=[1, 2]).distinct()\n<QuerySet [<Publication: Highlights for Children>, <Publication: Science News>, <Publication: Science Weekly>, <Publication: The Python Journal>]>\n>>> Publication.objects.filter(article__in=[a1, a2]).distinct()\n<QuerySet [<Publication: Highlights for Children>, <Publication: Science News>, <Publication: Science Weekly>, <Publication: The Python Journal>]>\nリレーション先のアイテムの除外も期待通りに動作します（SQLは少し複雑ですが）:
\n>>> Article.objects.exclude(publications=p2)\n<QuerySet [<Article: Django lets you build web apps easily>]>\nもし Publication を削除すると、その Articles にはアクセスできなくなります:
>>> p1.delete()\n>>> Publication.objects.all()\n<QuerySet [<Publication: Highlights for Children>, <Publication: Science News>, <Publication: Science Weekly>]>\n>>> a1 = Article.objects.get(pk=1)\n>>> a1.publications.all()\n<QuerySet []>\nもし Article を削除すると、その Publications にはアクセスできなくなります:
>>> a2.delete()\n>>> Article.objects.all()\n<QuerySet [<Article: Django lets you build web apps easily>]>\n>>> p2.article_set.all()\n<QuerySet []>\n多対多のリレーションシップの「反対側の」端を経由した追加:
\n>>> a4 = Article(headline="NASA finds intelligent life on Earth")\n>>> a4.save()\n>>> p2.article_set.add(a4)\n>>> p2.article_set.all()\n<QuerySet [<Article: NASA finds intelligent life on Earth>]>\n>>> a4.publications.all()\n<QuerySet [<Publication: Science News>]>\nキーワードを使った反対側の端からの追加:
\n>>> new_article = p2.article_set.create(headline="Oxygen-free diet works wonders")\n>>> p2.article_set.all()\n<QuerySet [<Article: NASA finds intelligent life on Earth>, <Article: Oxygen-free diet works wonders>]>\n>>> a5 = p2.article_set.all()[1]\n>>> a5.publications.all()\n<QuerySet [<Publication: Science News>]>\nArticle から Publication を削除する:
>>> a4.publications.remove(p2)\n>>> p2.article_set.all()\n<QuerySet [<Article: Oxygen-free diet works wonders>]>\n>>> a4.publications.all()\n<QuerySet []>\n反対側からも:
\n>>> p2.article_set.remove(a5)\n>>> p2.article_set.all()\n<QuerySet []>\n>>> a5.publications.all()\n<QuerySet []>\nリレーションのセットも指定できます:
\n>>> a4.publications.all()\n<QuerySet [<Publication: Science News>]>\n>>> a4.publications.set([p3])\n>>> a4.publications.all()\n<QuerySet [<Publication: Science Weekly>]>\nリレーションのセットを削除することもできます:
\n>>> p2.article_set.clear()\n>>> p2.article_set.all()\n<QuerySet []>\n反対側からも削除できます:
\n>>> p2.article_set.add(a4, a5)\n>>> p2.article_set.all()\n<QuerySet [<Article: NASA finds intelligent life on Earth>, <Article: Oxygen-free diet works wonders>]>\n>>> a4.publications.all()\n<QuerySet [<Publication: Science News>, <Publication: Science Weekly>]>\n>>> a4.publications.clear()\n>>> a4.publications.all()\n<QuerySet []>\n>>> p2.article_set.all()\n<QuerySet [<Article: Oxygen-free diet works wonders>]>\n削除した Article と Publication を再作成する:
>>> p1 = Publication(title="The Python Journal")\n>>> p1.save()\n>>> a2 = Article(headline="NASA uses Python")\n>>> a2.save()\n>>> a2.publications.add(p1, p2, p3)\n複数の Publications を一括削除 - 削除された出版物への参照は削除されます:
>>> Publication.objects.filter(title__startswith="Science").delete()\n>>> Publication.objects.all()\n<QuerySet [<Publication: Highlights for Children>, <Publication: The Python Journal>]>\n>>> Article.objects.all()\n<QuerySet [<Article: Django lets you build web apps easily>, <Article: NASA finds intelligent life on Earth>, <Article: NASA uses Python>, <Article: Oxygen-free diet works wonders>]>\n>>> a2.publications.all()\n<QuerySet [<Publication: The Python Journal>]>\n複数の記事を一括削除 - 削除したオブジェクトへの参照は削除されます:
\n>>> q = Article.objects.filter(headline__startswith="Django")\n>>> print(q)\n<QuerySet [<Article: Django lets you build web apps easily>]>\n>>> q.delete()\ndelete() の後は、 QuerySet のキャッシュをクリアする必要があります。参照されたオブジェクトは削除されます:
>>> print(q)\n<QuerySet []>\n>>> p1.article_set.all()\n<QuerySet [<Article: NASA uses Python>]>\n