Release: 1.4.0b1 pre release | Release Date: unreleased

SQLAlchemy 1.4 Documentation

Loading Columns

This section presents additional options regarding the loading of columns.

Deferred Column Loading

Deferred column loading allows particular columns of a table be loaded only upon direct access, instead of when the entity is queried using Query. This feature is useful when one wants to avoid loading a large text or binary field into memory when it’s not needed. Individual columns can be lazy loaded by themselves or placed into groups that lazy-load together, using the orm.deferred() function to mark them as “deferred”. In the example below, we define a mapping that will load each of .excerpt and .photo in separate, individual-row SELECT statements when each attribute is first referenced on the individual object instance:

from sqlalchemy.orm import deferred
from sqlalchemy import Integer, String, Text, Binary, Column

class Book(Base):
    __tablename__ = 'book'

    book_id = Column(Integer, primary_key=True)
    title = Column(String(200), nullable=False)
    summary = Column(String(2000))
    excerpt = deferred(Column(Text))
    photo = deferred(Column(Binary))

Classical mappings as always place the usage of orm.deferred() in the properties dictionary against the table-bound Column:

mapper(Book, book_table, properties={
    'photo':deferred(book_table.c.photo)
})

Deferred columns can be associated with a “group” name, so that they load together when any of them are first accessed. The example below defines a mapping with a photos deferred group. When one .photo is accessed, all three photos will be loaded in one SELECT statement. The .excerpt will be loaded separately when it is accessed:

class Book(Base):
    __tablename__ = 'book'

    book_id = Column(Integer, primary_key=True)
    title = Column(String(200), nullable=False)
    summary = Column(String(2000))
    excerpt = deferred(Column(Text))
    photo1 = deferred(Column(Binary), group='photos')
    photo2 = deferred(Column(Binary), group='photos')
    photo3 = deferred(Column(Binary), group='photos')

Deferred Column Loader Query Options

Columns can be marked as “deferred” or reset to “undeferred” at query time using options which are passed to the Query.options() method; the most basic query options are orm.defer() and orm.undefer():

from sqlalchemy.orm import defer
from sqlalchemy.orm import undefer

query = session.query(Book)
query = query.options(defer('summary'), undefer('excerpt'))
query.all()

Above, the “summary” column will not load until accessed, and the “excerpt” column will load immediately even if it was mapped as a “deferred” column.

orm.deferred() attributes which are marked with a “group” can be undeferred using orm.undefer_group(), sending in the group name:

from sqlalchemy.orm import undefer_group

query = session.query(Book)
query.options(undefer_group('photos')).all()

Deferred Loading across Multiple Entities

To specify column deferral for a Query that loads multiple types of entities at once, the deferral options may be specified more explicitly using class-bound attributes, rather than string names:

from sqlalchemy.orm import defer

query = session.query(Book, Author).join(Book.author)
query = query.options(defer(Author.bio))

Column deferral options may also indicate that they take place along various relationship paths, which are themselves often eagerly loaded with loader options. All relationship-bound loader options support chaining onto additional loader options, which include loading for further levels of relationships, as well as onto column-oriented attributes at that path. Such as, to load Author instances, then joined-eager-load the Author.books collection for each author, then apply deferral options to column-oriented attributes onto each Book entity from that relationship, the joinedload() loader option can be combined with the load_only() option (described later in this section) to defer all Book columns except those explicitly specified:

from sqlalchemy.orm import joinedload

query = session.query(Author)
query = query.options(
            joinedload(Author.books).load_only(Book.summary, Book.excerpt),
        )

Option structures as above can also be organized in more complex ways, such as hierarchically using the Load.options() method, which allows multiple sub-options to be chained to a common parent option at once. Any mixture of string names and class-bound attribute objects may be used:

from sqlalchemy.orm import defer
from sqlalchemy.orm import joinedload
from sqlalchemy.orm import load_only

query = session.query(Author)
query = query.options(
            joinedload(Author.book).options(
                load_only("summary", "excerpt"),
                joinedload(Book.citations).options(
                    joinedload(Citation.author),
                    defer(Citation.fulltext)
                )
            )
        )

New in version 1.3.6: Added Load.options() to allow easier construction of hierarchies of loader options.

Another way to apply options to a path is to use the orm.defaultload() function. This function is used to indicate a particular path within a loader option structure without actually setting any options at that level, so that further sub-options may be applied. The orm.defaultload() function can be used to create the same structure as we did above using Load.options() as:

query = session.query(Author)
query = query.options(
    joinedload(Author.book).load_only("summary", "excerpt"),
    defaultload(Author.book).joinedload(Book.citations).joinedload(Citation.author),
    defaultload(Author.book).defaultload(Book.citations).defer(Citation.fulltext)
)

See also

Relationship Loading with Loader Options - targeted towards relationship loading

Load Only and Wildcard Options

The ORM loader option system supports the concept of “wildcard” loader options, in which a loader option can be passed an asterisk "*" to indicate that a particular option should apply to all applicable attributes of a mapped class. Such as, if we wanted to load the Book class but only the “summary” and “excerpt” columns, we could say:

from sqlalchemy.orm import defer
from sqlalchemy.orm import undefer

session.query(Book).options(
    defer('*'), undefer("summary"), undefer("excerpt"))

Above, the defer() option is applied using a wildcard to all column attributes on the Book class. Then, the undefer() option is used against the “summary” and “excerpt” fields so that they are the only columns loaded up front. A query for the above entity will include only the “summary” and “excerpt” fields in the SELECT, along with the primary key columns which are always used by the ORM.

A similar function is available with less verbosity by using the orm.load_only() option. This is a so-called exclusionary option which will apply deferred behavior to all column attributes except those that are named:

from sqlalchemy.orm import load_only

session.query(Book).options(load_only("summary", "excerpt"))

Wildcard and Exclusionary Options with Multiple-Entity Queries

Wildcard options and exclusionary options such as load_only() may only be applied to a single entity at a time within a Query. To suit the less common case where a Query is returning multiple primary entities at once, a special calling style may be required in order to apply a wildcard or exclusionary option, which is to use the Load object to indicate the starting entity for a deferral option. Such as, if we were loading Book and Author at once, the Query will raise an informative error if we try to apply load_only() to both at once. Using Load looks like:

from sqlalchemy.orm import Load

query = session.query(Book, Author).join(Book.author)
query = query.options(
            Load(Book).load_only("summary", "excerpt")
        )

Above, Load is used in conjunction with the exclusionary option load_only() so that the deferral of all other columns only takes place for the Book class and not the Author class. Again, the Query object should raise an informative error message when the above calling style is actually required that describes those cases where explicit use of Load is needed.

Raiseload for Deferred Columns

New in version 1.4.

The deferred() loader option and the corresponding loader strategy also support the concept of “raiseload”, which is a loader strategy that will raise InvalidRequestError if the attribute is accessed such that it would need to emit a SQL query in order to be loaded. This behavior is the column-based equivalent of the raiseload() feature for relationship loading, discussed at Preventing unwanted lazy loads using raiseload. Using the orm.defer.raiseload parameter on the defer() option, an exception is raised if the attribute is accessed:

book = session.query(Book).options(defer(Book.summary, raiseload=True)).first()

# would raise an exception
book.summary

Deferred “raiseload” can be configured at the mapper level via orm.deferred.raiseload on deferred(), so that an explicit undefer() is required in order for the attribute to be usable:

class Book(Base):
    __tablename__ = 'book'

    book_id = Column(Integer, primary_key=True)
    title = Column(String(200), nullable=False)
    summary = deferred(Column(String(2000)), raiseload=True)
    excerpt = deferred(Column(Text), raiseload=True)

book_w_excerpt = session.query(Book).options(undefer(Book.excerpt)).first()

Column Deferral API

sqlalchemy.orm.defer(key, *addl_attrs, **kw)

Indicate that the given column-oriented attribute should be deferred, e.g. not loaded until accessed.

This function is part of the Load interface and supports both method-chained and standalone operation.

e.g.:

from sqlalchemy.orm import defer

session.query(MyClass).options(
                    defer("attribute_one"),
                    defer("attribute_two"))

session.query(MyClass).options(
                    defer(MyClass.attribute_one),
                    defer(MyClass.attribute_two))

To specify a deferred load of an attribute on a related class, the path can be specified one token at a time, specifying the loading style for each link along the chain. To leave the loading style for a link unchanged, use orm.defaultload():

session.query(MyClass).options(defaultload("someattr").defer("some_column"))

A Load object that is present on a certain path can have Load.defer() called multiple times, each will operate on the same parent entity:

session.query(MyClass).options(
                defaultload("someattr").
                    defer("some_column").
                    defer("some_other_column").
                    defer("another_column")
    )
Parameters
  • key – Attribute to be deferred.

  • raiseload

    raise InvalidRequestError if the column value is to be loaded from emitting SQL. Used to prevent unwanted SQL from being emitted.

    New in version 1.4.

  • *addl_attrs

    This option supports the old 0.8 style of specifying a path as a series of attributes, which is now superseded by the method-chained style.

    Deprecated since version 0.9: The *addl_attrs on orm.defer() is deprecated and will be removed in a future release. Please use method chaining in conjunction with defaultload() to indicate a path.

sqlalchemy.orm.deferred(*columns, **kw)

Indicate a column-based mapped attribute that by default will not load unless accessed.

Parameters
  • *columns – columns to be mapped. This is typically a single Column object, however a collection is supported in order to support multiple columns mapped under the same attribute.

  • raiseload

    boolean, if True, indicates an exception should be raised if the load operation is to take place.

    New in version 1.4.

  • **kw – additional keyword arguments passed to ColumnProperty.

sqlalchemy.orm.query_expression()

Indicate an attribute that populates from a query-time SQL expression.

New in version 1.2.

See also

mapper_query_expression

sqlalchemy.orm.load_only(*attrs)

Indicate that for a particular entity, only the given list of column-based attribute names should be loaded; all others will be deferred.

This function is part of the Load interface and supports both method-chained and standalone operation.

Example - given a class User, load only the name and fullname attributes:

session.query(User).options(load_only("name", "fullname"))

Example - given a relationship User.addresses -> Address, specify subquery loading for the User.addresses collection, but on each Address object load only the email_address attribute:

session.query(User).options(
        subqueryload("addresses").load_only("email_address")
)

For a Query that has multiple entities, the lead entity can be specifically referred to using the Load constructor:

session.query(User, Address).join(User.addresses).options(
            Load(User).load_only("name", "fullname"),
            Load(Address).load_only("email_addres")
        )

New in version 0.9.0.

sqlalchemy.orm.undefer(key, *addl_attrs)

Indicate that the given column-oriented attribute should be undeferred, e.g. specified within the SELECT statement of the entity as a whole.

The column being undeferred is typically set up on the mapping as a deferred() attribute.

This function is part of the Load interface and supports both method-chained and standalone operation.

Examples:

# undefer two columns
session.query(MyClass).options(undefer("col1"), undefer("col2"))

# undefer all columns specific to a single class using Load + *
session.query(MyClass, MyOtherClass).options(
    Load(MyClass).undefer("*"))

# undefer a column on a related object
session.query(MyClass).options(
    defaultload(MyClass.items).undefer('text'))
Parameters
  • key – Attribute to be undeferred.

  • *addl_attrs

    This option supports the old 0.8 style of specifying a path as a series of attributes, which is now superseded by the method-chained style.

    Deprecated since version 0.9: The *addl_attrs on orm.undefer() is deprecated and will be removed in a future release. Please use method chaining in conjunction with defaultload() to indicate a path.

sqlalchemy.orm.undefer_group(name)

Indicate that columns within the given deferred group name should be undeferred.

The columns being undeferred are set up on the mapping as deferred() attributes and include a “group” name.

E.g:

session.query(MyClass).options(undefer_group("large_attrs"))

To undefer a group of attributes on a related entity, the path can be spelled out using relationship loader options, such as orm.defaultload():

session.query(MyClass).options(
    defaultload("someattr").undefer_group("large_attrs"))

Changed in version 0.9.0: orm.undefer_group() is now specific to a particular entity load path.

sqlalchemy.orm.with_expression(key, expression)

Apply an ad-hoc SQL expression to a “deferred expression” attribute.

This option is used in conjunction with the orm.query_expression() mapper-level construct that indicates an attribute which should be the target of an ad-hoc SQL expression.

E.g.:

sess.query(SomeClass).options(
    with_expression(SomeClass.x_y_expr, SomeClass.x + SomeClass.y)
)

New in version 1.2.

Parameters
  • key – Attribute to be undeferred.

  • expr – SQL expression to be applied to the attribute.

See also

mapper_query_expression

Column Bundles

The Bundle may be used to query for groups of columns under one namespace.

New in version 0.9.0.

The bundle allows columns to be grouped together:

from sqlalchemy.orm import Bundle

bn = Bundle('mybundle', MyClass.data1, MyClass.data2)
for row in session.query(bn).filter(bn.c.data1 == 'd1'):
    print(row.mybundle.data1, row.mybundle.data2)

The bundle can be subclassed to provide custom behaviors when results are fetched. The method Bundle.create_row_processor() is given the Query and a set of “row processor” functions at query execution time; these processor functions when given a result row will return the individual attribute value, which can then be adapted into any kind of return data structure. Below illustrates replacing the usual KeyedTuple return structure with a straight Python dictionary:

from sqlalchemy.orm import Bundle

class DictBundle(Bundle):
    def create_row_processor(self, query, procs, labels):
        """Override create_row_processor to return values as dictionaries"""
        def proc(row):
            return dict(
                        zip(labels, (proc(row) for proc in procs))
                    )
        return proc

Changed in version 1.0: The proc() callable passed to the create_row_processor() method of custom Bundle classes now accepts only a single “row” argument.

A result from the above bundle will return dictionary values:

bn = DictBundle('mybundle', MyClass.data1, MyClass.data2)
for row in session.query(bn).filter(bn.c.data1 == 'd1'):
    print(row.mybundle['data1'], row.mybundle['data2'])

The Bundle construct is also integrated into the behavior of composite(), where it is used to return composite attributes as objects when queried as individual attributes.

Previous: Loading Objects Next: Relationship Loading Techniques