Special Relationship Persistence Patterns

Rows that point to themselves / Mutually Dependent Rows

This is a very specific case where relationship() must perform an INSERT and a second UPDATE in order to properly populate a row (and vice versa an UPDATE and DELETE in order to delete without violating foreign key constraints). The two use cases are:

  • A table contains a foreign key to itself, and a single row will have a foreign key value pointing to its own primary key.

  • Two tables each contain a foreign key referencing the other table, with a row in each table referencing the other.

For example:

          user
---------------------------------
user_id    name   related_user_id
   1       'ed'          1

Or:

             widget                                                  entry
-------------------------------------------             ---------------------------------
widget_id     name        favorite_entry_id             entry_id      name      widget_id
   1       'somewidget'          5                         5       'someentry'     1

In the first case, a row points to itself. Technically, a database that uses sequences such as PostgreSQL or Oracle can INSERT the row at once using a previously generated value, but databases which rely upon autoincrement-style primary key identifiers cannot. The relationship() always assumes a “parent/child” model of row population during flush, so unless you are populating the primary key/foreign key columns directly, relationship() needs to use two statements.

In the second case, the “widget” row must be inserted before any referring “entry” rows, but then the “favorite_entry_id” column of that “widget” row cannot be set until the “entry” rows have been generated. In this case, it’s typically impossible to insert the “widget” and “entry” rows using just two INSERT statements; an UPDATE must be performed in order to keep foreign key constraints fulfilled. The exception is if the foreign keys are configured as “deferred until commit” (a feature some databases support) and if the identifiers were populated manually (again essentially bypassing relationship()).

To enable the usage of a supplementary UPDATE statement, we use the relationship.post_update option of relationship(). This specifies that the linkage between the two rows should be created using an UPDATE statement after both rows have been INSERTED; it also causes the rows to be de-associated with each other via UPDATE before a DELETE is emitted. The flag should be placed on just one of the relationships, preferably the many-to-one side. Below we illustrate a complete example, including two ForeignKey constructs:

from sqlalchemy import Integer, ForeignKey, Column
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import relationship

Base = declarative_base()

class Entry(Base):
    __tablename__ = 'entry'
    entry_id = Column(Integer, primary_key=True)
    widget_id = Column(Integer, ForeignKey('widget.widget_id'))
    name = Column(String(50))

class Widget(Base):
    __tablename__ = 'widget'

    widget_id = Column(Integer, primary_key=True)
    favorite_entry_id = Column(Integer,
                            ForeignKey('entry.entry_id',
                            name="fk_favorite_entry"))
    name = Column(String(50))

    entries = relationship(Entry, primaryjoin=
                                    widget_id==Entry.widget_id)
    favorite_entry = relationship(Entry,
                                primaryjoin=
                                    favorite_entry_id==Entry.entry_id,
                                post_update=True)

When a structure against the above configuration is flushed, the “widget” row will be INSERTed minus the “favorite_entry_id” value, then all the “entry” rows will be INSERTed referencing the parent “widget” row, and then an UPDATE statement will populate the “favorite_entry_id” column of the “widget” table (it’s one row at a time for the time being):

>>> w1 = Widget(name='somewidget')
>>> e1 = Entry(name='someentry')
>>> w1.favorite_entry = e1
>>> w1.entries = [e1]
>>> session.add_all([w1, e1])
sql>>> session.commit()

An additional configuration we can specify is to supply a more comprehensive foreign key constraint on Widget, such that it’s guaranteed that favorite_entry_id refers to an Entry that also refers to this Widget. We can use a composite foreign key, as illustrated below:

from sqlalchemy import Integer, ForeignKey, String, \
        Column, UniqueConstraint, ForeignKeyConstraint
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import relationship

Base = declarative_base()

class Entry(Base):
    __tablename__ = 'entry'
    entry_id = Column(Integer, primary_key=True)
    widget_id = Column(Integer, ForeignKey('widget.widget_id'))
    name = Column(String(50))
    __table_args__ = (
        UniqueConstraint("entry_id", "widget_id"),
    )

class Widget(Base):
    __tablename__ = 'widget'

    widget_id = Column(Integer, autoincrement='ignore_fk', primary_key=True)
    favorite_entry_id = Column(Integer)

    name = Column(String(50))

    __table_args__ = (
        ForeignKeyConstraint(
            ["widget_id", "favorite_entry_id"],
            ["entry.widget_id", "entry.entry_id"],
            name="fk_favorite_entry"
        ),
    )

    entries = relationship(Entry, primaryjoin=
                                    widget_id==Entry.widget_id,
                                    foreign_keys=Entry.widget_id)
    favorite_entry = relationship(Entry,
                                primaryjoin=
                                    favorite_entry_id==Entry.entry_id,
                                foreign_keys=favorite_entry_id,
                                post_update=True)

The above mapping features a composite ForeignKeyConstraint bridging the widget_id and favorite_entry_id columns. To ensure that Widget.widget_id remains an “autoincrementing” column we specify Column.autoincrement to the value "ignore_fk" on Column, and additionally on each relationship() we must limit those columns considered as part of the foreign key for the purposes of joining and cross-population.

Mutable Primary Keys / Update Cascades

When the primary key of an entity changes, related items which reference the primary key must also be updated as well. For databases which enforce referential integrity, the best strategy is to use the database’s ON UPDATE CASCADE functionality in order to propagate primary key changes to referenced foreign keys - the values cannot be out of sync for any moment unless the constraints are marked as “deferrable”, that is, not enforced until the transaction completes.

It is highly recommended that an application which seeks to employ natural primary keys with mutable values to use the ON UPDATE CASCADE capabilities of the database. An example mapping which illustrates this is:

class User(Base):
    __tablename__ = 'user'
    __table_args__ = {'mysql_engine': 'InnoDB'}

    username = Column(String(50), primary_key=True)
    fullname = Column(String(100))

    addresses = relationship("Address")


class Address(Base):
    __tablename__ = 'address'
    __table_args__ = {'mysql_engine': 'InnoDB'}

    email = Column(String(50), primary_key=True)
    username = Column(String(50),
                ForeignKey('user.username', onupdate="cascade")
            )

Above, we illustrate onupdate="cascade" on the ForeignKey object, and we also illustrate the mysql_engine='InnoDB' setting which, on a MySQL backend, ensures that the InnoDB engine supporting referential integrity is used. When using SQLite, referential integrity should be enabled, using the configuration described at Foreign Key Support.

See also

Using foreign key ON DELETE cascade with ORM relationships - supporting ON DELETE CASCADE with relationships

mapper.passive_updates - similar feature on mapper()

Simulating limited ON UPDATE CASCADE without foreign key support

In those cases when a database that does not support referential integrity is used, and natural primary keys with mutable values are in play, SQLAlchemy offers a feature in order to allow propagation of primary key values to already-referenced foreign keys to a limited extent, by emitting an UPDATE statement against foreign key columns that immediately reference a primary key column whose value has changed. The primary platforms without referential integrity features are MySQL when the MyISAM storage engine is used, and SQLite when the PRAGMA foreign_keys=ON pragma is not used. The Oracle database also has no support for ON UPDATE CASCADE, but because it still enforces referential integrity, needs constraints to be marked as deferrable so that SQLAlchemy can emit UPDATE statements.

The feature is enabled by setting the relationship.passive_updates flag to False, most preferably on a one-to-many or many-to-many relationship(). When “updates” are no longer “passive” this indicates that SQLAlchemy will issue UPDATE statements individually for objects referenced in the collection referred to by the parent object with a changing primary key value. This also implies that collections will be fully loaded into memory if not already locally present.

Our previous mapping using passive_updates=False looks like:

class User(Base):
    __tablename__ = 'user'

    username = Column(String(50), primary_key=True)
    fullname = Column(String(100))

    # passive_updates=False *only* needed if the database
    # does not implement ON UPDATE CASCADE
    addresses = relationship("Address", passive_updates=False)

class Address(Base):
    __tablename__ = 'address'

    email = Column(String(50), primary_key=True)
    username = Column(String(50), ForeignKey('user.username'))

Key limitations of passive_updates=False include:

  • it performs much more poorly than direct database ON UPDATE CASCADE, because it needs to fully pre-load affected collections using SELECT and also must emit UPDATE statements against those values, which it will attempt to run in “batches” but still runs on a per-row basis at the DBAPI level.

  • the feature cannot “cascade” more than one level. That is, if mapping X has a foreign key which refers to the primary key of mapping Y, but then mapping Y’s primary key is itself a foreign key to mapping Z, passive_updates=False cannot cascade a change in primary key value from Z to X.

  • Configuring passive_updates=False only on the many-to-one side of a relationship will not have a full effect, as the unit of work searches only through the current identity map for objects that may be referencing the one with a mutating primary key, not throughout the database.

As virtually all databases other than Oracle now support ON UPDATE CASCADE, it is highly recommended that traditional ON UPDATE CASCADE support be used in the case that natural and mutable primary key values are in use.